A method and system for automatically extracting bridge edge lines from point clouds

Abstract

This invention relates to the field of engineering surveying technology, specifically to a method and system for automatically extracting bridge edge lines from point clouds. The method includes the following steps: acquiring a 3D point cloud dataset of the bridge and its surroundings; determining initial edge line points, filtering the point cloud based on a circular neighborhood and converting to polar coordinates; eliminating interference points through elevation change detection; calculating the initial normal azimuth angle; locating the initial edge line points on the opposite side of the bridge; iteratively searching for adjacent edge line points along a preset bridge direction, using the initial edge line points on both sides of the bridge as seed points; and determining the termination condition of the search through an included angle threshold to generate complete double-sided edge lines of the bridge. The system includes modules for data input, initial settings, direction calculation, opposite-side positioning, iterative detection, and result output. This invention solves the problems of high safety risks and low efficiency in traditional manual surveying, as well as the poor accuracy of point cloud edge line extraction by manual vision. It achieves automated and high-precision extraction of bridge edge lines in complex environments, providing intuitive quantitative basis for construction decisions.

Description

Technical Field

[0001] This invention relates to the field of bridge engineering surveying and three-dimensional data processing technology, and in particular to a method and system for automatically extracting bridge edge lines from point clouds. Background Technology

[0002] In the construction of infrastructure such as railways and highways, bridge boundary line extraction is a crucial preliminary step for achieving track safety assessment, clearance detection, and digital modeling. This is especially important when the construction project is adjacent to or intersects with an existing bridge. To ensure construction safety and structural compliance, it is essential to accurately detect the boundary line position of the existing bridge to determine its spatial relationship with the new project, providing a basis for construction decisions.

[0003] Existing methods for bridge edge detection include manual measurement and manual extraction based on point cloud data, which mainly have the following problems:

[0004] Manual surveying is high-risk and inefficient. Traditional methods require surveyors to carry instruments and collect coordinates point by point along the bridge's edge. For high-pier, long-span bridges or bridges spanning rivers or canyons, workers must work close to the edge or even suspended in mid-air, posing a significant risk of falls from heights. Furthermore, manual point-by-point surveying is time-consuming, labor-intensive, and inefficient.

[0005] The accuracy of manual extraction of point cloud data is poor. Although 3D laser scanning technology is now widespread and can quickly acquire bridge point clouds, current point cloud processing largely relies on manual visual identification of edges point by point in software. Faced with millions or even tens of millions of point cloud data, manual edge identification is not only labor-intensive but also highly susceptible to subjective factors, easily leading to the misselection of ground points, vegetation points, or non-edge points in the middle of the bridge. This results in low accuracy of the extracted edges, making it difficult to meet the needs of refined construction.

[0006] Therefore, developing a method and system that can automatically, accurately, and safely extract bridge edge lines using point cloud data has significant engineering application value. Summary of the Invention

[0007] The purpose of this invention is to overcome the problems of high risk, low efficiency and insufficient accuracy of manual measurement and point cloud edge line extraction in the prior art, and to provide a method and system for automatically extracting bridge edge lines from point clouds.

[0008] In a first aspect, the present invention provides a method for automatically extracting bridge edge lines from a point cloud, the steps of which include: S1. Obtain a 3D point cloud dataset containing the bridge.

[0009] Preferably, any point in the 3D point cloud dataset is determined by the coordinates of the construction coordinate system. Unique identifier, among which For the north coordinate component, For the eastern coordinate component, This represents the elevation component.

[0010] S2. Determine the initial edge point Q in the point cloud dataset; with Q as the center, Inner diameter A ring-shaped neighborhood filtering region is constructed for the outer diameter, filtering all neighborhood points that satisfy the constraints; with Q as the origin of the polar coordinates, the neighborhood points are transformed into polar coordinates and a neighborhood dataset is constructed, represented as:

[0011] In the formula, P represents the neighborhood dataset. This represents the horizontal distance from point i to Q. This represents the azimuth angle of Q pointing to the i-th point. Indicate the elevation of point i; Preferably, the initial edge point Q is located in the edge area of ​​the bridge beam and maintains a preset safe distance from the pier structure; its spatial coordinates are obtained through on-site measurement or through manual interactive extraction based on the point cloud data.

[0012] Preferably, the constraint condition is:

[0013] Where r is the horizontal distance from the neighboring point to the initial edge point Q; inner diameter The value is negatively correlated with the spatial distribution density of the edge points to be extracted; outer diameter The value is less than the horizontal distance from Q to the nearest pier.

[0014] S3. Detect elevation abrupt changes using a preset method to remove non-bridge points and obtain a clean bridge point dataset P1.

[0015] Preferably, the preset method for detecting abrupt changes in elevation is as follows: S31. Based on the stated elevation The neighborhood dataset P is sorted in descending order of size. S32. Traverse and calculate the elevation difference between adjacent data. ; S33. Based on the preset elevation change threshold Judgment, if If the (i+1)th point is an elevation change point, then the (i+1)th element and its subsequent elements are non-bridge points. S34. Record the current critical elevation h0= The dataset is truncated, and the first to the ith elements are retained to obtain the pure bridge point dataset P1. The threshold Set to the minimum height difference between the bottom of the bridge and the ground.

[0016] S4. Based on the pure bridge point dataset P1, detect and calculate the initial normal azimuth A0 and the reverse azimuth A1.

[0017] Preferably, the initial normal azimuth angle A0 is perpendicular to the bridge's orientation and points outwards from the bridge; the reverse azimuth angle A1 is perpendicular to the bridge's orientation and points inwards from the bridge; the detection calculation process is as follows: S41. Arrange the pure bridge point dataset P1 according to coordinate azimuth angle Sort in ascending order; perform a data loop operation, copying the first element of the sorted dataset and appending it to the end of the dataset to construct periodic boundary conditions; S42. Traverse the dataset after loop closure and calculate the azimuth increment of adjacent elements. ,like Then the azimuth increment is corrected to ; S43. Extract the maximum value from all azimuth increments. Second largest value And its corresponding indices m and n, comparing the maximum and second largest values ​​relative to each other. Degree of deviation: like , If the absolute value is taken, the direction of the maximum gap is determined to be the direction of the bridge's external normal. The calculation formula is as follows:

[0018] Otherwise, the direction of the second largest gap is determined to be the direction of the bridge's external normal, and the calculation formula is:

[0019] Normalize the calculated A0. Then it is corrected to ; S44. Calculate the reverse azimuth angle ,like Then it is corrected to .

[0020] S5. Based on the initial edge point Q and the reverse azimuth angle A1, locate the initial edge point W on the opposite side of the bridge.

[0021] Preferably, the positioning process of the initial edge point W on the opposite side of the bridge includes: S51. Traverse the neighborhood dataset P, based on the azimuth angle. Constraints are used to filter the dataset, removing those that do not meet the angle range. The interference point has the following azimuth constraint condition:

[0022] In the formula, α is a preset angle range, and its value is negatively correlated with the width of the bridge; S52. Traverse the dataset after the removal operation in S51, based on elevation... Constraints are used to filter the dataset, removing data that does not meet the elevation threshold. The interference points have the following elevation constraints:

[0023] In the formula, This is the critical elevation. S53. Calculate the point with the maximum horizontal distance from each of the remaining points in the dataset to the initial edge point Q, and use it as the initial edge point W on the opposite side of the bridge.

[0024] S6. Using the initial edge point Q and the opposite initial edge point W as starting seed points respectively, based on the azimuth angle A0 and the reverse azimuth angle A1, an iterative stepping strategy is adopted to search for adjacent edge points along both sides of the bridge longitudinal direction. The change in the angle between the current search point and the preceding and following adjacent points is calculated in real time, and the search termination condition is determined by the angle threshold. Finally, a complete array of bridge double-sided edge point sets is generated.

[0025] Preferably, the retrieval steps for adjacent edge points are as follows: S61. Traverse the neighborhood dataset P, initially setting the current node as the initial edge point Q, and so on. , Centered on the initial normal azimuth angle A0, calculate the clockwise angle between the line connecting the current point and Q and A0, and select the point with the minimum angle as the right extension point of A0. Select the point with the maximum included angle as the left extension point of A0. This enables traversal and retrieval of edge points along both sides of the bridge's longitudinal direction. S62. Starting from the initial edge point W on the opposite side of the bridge, repeat the retrieval steps based on the reverse azimuth angle A1 to extract the edge point on the other side of the bridge; S63. When the change in the included angle exceeds a preset included angle threshold, it is determined that the bridge range has been exceeded and the search is stopped.

[0026] Preferably, the included angle threshold γ is preset to be .

[0027] In a second aspect, the present invention provides a system for automatically extracting bridge edge lines from point clouds, comprising: The data input module is used to load and acquire 3D point cloud data; The initial setup module is used to select initial edge points and set algorithm parameters in the point cloud view; The orientation calculation module is used to perform polar coordinate transformation, elevation change detection, and automatic calculation of the initial normal azimuth angle; The opposite side positioning module is used to determine the starting edge point on the opposite side of the bridge; The iterative detection module is used to execute the edge point retrieval algorithm. The results output module is used to generate and store the bridge edge point array; Preferably, the system can run on an engineering surveying terminal or a cloud server, supports interface docking with a construction management platform, and directly uses the extracted boundary data for spatial positional relationship analysis between new construction projects and existing bridges.

[0028] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention provides a method for automatically extracting bridge edges from point clouds, which automatically completes neighborhood search, noise removal, direction calculation and point tracking. Compared with manual visual point picking, the efficiency is improved by several times and the degree of automation is high.

[0029] 2. This invention provides a method for automatically extracting bridge edges from point clouds. It effectively separates bridge surface points from ground / vegetation points by using elevation change detection, avoiding false extractions. Through a circular neighborhood iteration mechanism, it sensitively detects minute changes in bridge edges. It adopts angle change detection as a termination retrieval condition to intelligently identify bridge endpoints and prevent the algorithm from overflowing onto approach bridges or roadbeds. The extraction accuracy is high and the robustness is strong.

[0030] 3. This invention provides a method for automatically extracting bridge edges from point clouds, accurately quantifying the relationships such as clearance, intersection angle, and spatial avoidance between new construction projects and existing bridges, effectively avoiding construction collision risks, and ensuring the operational safety of existing bridges and the construction safety of new projects; based on existing point cloud data, there is no need for on-site measurements by construction personnel, completely eliminating the safety hazards of high-altitude operations and improving the safety of construction personnel; it provides an intuitive, reliable, and quantifiable data barrier for construction in complex environments.

[0031] 4. This invention provides a system for automatically extracting bridge edges from point clouds. By adjusting the key core parameters of the system, it can be directly transferred to different engineering projects, adapting to various bridge structures with different widths, heights and point cloud densities. It is highly adaptable and saves resources. Attached Figure Description

[0032] Figure 1 Flowchart of a method for automatically extracting bridge edges from point clouds; Figure 2This is a schematic diagram illustrating the principle of circular neighborhood filtering. Figure 3 This is a schematic diagram showing the distribution of clean bridge points after elevation change detection; Figure 4 This is a schematic diagram of filtering a neighborhood dataset based on azimuth constraints. Figure 5 Flowchart for traversing and retrieving edge points; Figure 6 A schematic diagram illustrating the principle of traversing and retrieving edge points; Figure 7 A schematic diagram illustrating the termination determination of iterative retrieval of edge points; Figure 8 This is the original point cloud data of a bridge and its surrounding environment. Figure 9 This is a 2D point cloud view of a bridge. Figure 10 This is a diagram showing the extracted bridge edge lines. Detailed Implementation

[0033] The present invention will now be described in further detail with reference to specific embodiments. However, this should not be construed as limiting the scope of the present invention to the following embodiments; all technologies implemented based on the content of the present invention fall within the scope of the present invention.

[0034] Unless otherwise specified, the use of terms such as "upper," "lower," "left," "right," "center," "inner," and "outer" to indicate orientation or positional relationships in the description of specific embodiments of the present invention is based on the orientation or positional relationships shown in the accompanying drawings, or the orientation or positional relationship in which the product / equipment / device is typically placed during use. These terms are merely for the purpose of facilitating the description of the present invention or simplifying the description in specific embodiments, and for enabling those skilled in the art to quickly understand the solution. They do not indicate or imply that a particular device / component / element must have a specific orientation, or be constructed and operated in a specific positional relationship, and therefore should not be construed as limiting the present invention.

[0035] Furthermore, the use of terms such as "horizontal," "vertical," "suspended," "parallel," and "coaxial" does not imply that the corresponding device / component / element must be absolutely horizontal, vertical, suspended, parallel, or coaxial. Slight tilt or deviation is permissible, as long as it does not affect the normal function of the relevant component. For example, "horizontal" simply means that its direction is more horizontal relative to "vertical," not that the structure must be perfectly horizontal; a slight tilt is acceptable. "Coaxial" means that two components are arranged as coaxially as possible, allowing them to move coaxially or approximately coaxially when their relative positions change. Alternatively, it can be simplified to mean that the corresponding device / component / element, when arranged in "horizontal," "vertical," "suspended," "parallel," or "coaxial" directions, can have an error / deviation of ±10% relative to the corresponding direction, more preferably within ±8%, more preferably within ±6%, more preferably within ±5%, and more preferably within ±4%. For example, the deviation in the "coaxial" direction is controlled within 0.2-1mm, preferably within 0.2-0.5mm. As long as the corresponding device / component / element is within the error / deviation range, it can still achieve its function in the solution of the present invention.

[0036] Furthermore, the use of terms such as "first," "second," and "third" in terminology is merely for distinguishing descriptions of identical or similar components and should not be interpreted as emphasizing or implying the relative importance of a particular component.

[0037] Furthermore, in the description of the embodiments of the present invention, "several", "more than", and "a number of" represent at least two. The number can be any number, such as two, three, four, five, six, seven, eight, or nine, and can even exceed nine.

[0038] Furthermore, in the description of the technical solution of this invention, unless otherwise explicitly specified / limited / restricted, the terms "set up," "install," "connect," "link," "provided with," "laid out," and "arranged" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to connection methods commonly used in the art, such as welding, riveting, bolting, and threaded connections. Such connections can be mechanical, electrical, or communication connections; they can be direct connections or indirect connections through an intermediate medium; and they can refer to the internal communication between two components.

[0039] Example 1 This embodiment provides a method for automatically extracting bridge edge lines from point clouds, the flowchart of which is shown below. Figure 1 As shown, the steps include: S1. Obtain a 3D point cloud dataset containing the bridge.

[0040] Preferably, any point in the 3D point cloud dataset is determined by the coordinates of the construction coordinate system. Unique identifier, among which For the north coordinate component, For the eastern coordinate component, This represents the elevation component.

[0041] S2. Determine the initial edge point Q in the point cloud dataset; with Q as the center, Inner diameter A ring-shaped neighborhood filtering region is constructed for the outer diameter, filtering all neighborhood points that satisfy the constraints; with Q as the origin of the polar coordinates, the neighborhood points are transformed into polar coordinates and a neighborhood dataset is constructed, represented as:

[0042] In the formula, P represents the neighborhood dataset. This represents the horizontal distance from point i to Q. This represents the azimuth angle of Q pointing to the i-th point. Indicate the elevation of point i; Preferably, the initial edge point Q is located in the edge area of ​​the bridge beam and maintains a preset safe distance from the pier structure; its spatial coordinates are obtained through on-site measurement or through manual interactive extraction based on the point cloud data.

[0043] Preferably, the constraint conditions of the annular neighborhood screening region are:

[0044] Where r is the horizontal distance from the neighboring point to the initial edge point Q; inner diameter The value of R1 is negatively correlated with the spatial distribution density of the edge points to be extracted; that is, the higher the distribution density of the edge points to be extracted, the smaller the value of R1. outer diameter The value is less than the horizontal distance from Q to the nearest pier.

[0045] A schematic diagram of the circular neighborhood filtering principle is shown below. Figure 2 As shown.

[0046] S3. Detect elevation abrupt changes using a preset method to remove non-bridge points and obtain a clean bridge point dataset P1.

[0047] Based on the continuity characteristic of railway bridge structures, if a sharp change in elevation value occurs in the point cloud data along its extension direction (i.e., an elevation abrupt change), it indicates the termination of the main bridge structure. The starting point of the elevation abrupt change and all subsequent point cloud data should be identified as invalid non-bridge point data, such as those corresponding to the ground or vegetation under the bridge. Preferably, the preset method for detecting elevation abrupt changes is as follows: S31. Based on the stated elevation The neighborhood dataset P is sorted in descending order of size; S32. Traverse and calculate the elevation difference between adjacent data. ; S33. Based on the preset elevation change threshold Judgment, if ， Then the (i+1)th point is an elevation change point, and the (i+1)th element and its subsequent elements are non-bridge points; S34. Record the current critical elevation h0= The dataset is truncated, and the first to the ith elements are retained to obtain the pure bridge point dataset P1. The threshold Set to the minimum elevation difference between the bottom of the bridge and the ground. After step S3, invalid data is removed, and only clean point cloud data identified as bridge areas are retained, such as... Figure 3 As shown, the points in the pure bridge point dataset P1 exhibit an approximately semi-circular distribution.

[0048] S4. Based on the pure bridge point dataset P1, detect and calculate the initial normal azimuth A0 and the reverse azimuth A1.

[0049] Preferably, the initial normal azimuth angle A0 is perpendicular to the bridge's orientation and points outwards from the bridge; the reverse azimuth angle A1 is perpendicular to the bridge's orientation and points inwards from the bridge; the detection calculation process is as follows: S41. Arrange the pure bridge point dataset P1 according to coordinate azimuth angle Sort in ascending order; perform a data loop operation, copying the first element of the sorted dataset and appending it to the end of the dataset to construct periodic boundary conditions; This step establishes the spatial adjacency topology of point cloud data along the circumference by rearranging the polar coordinate data in ascending order of azimuth angle. This solves the problem of spatial disorder caused by sorting the original data by elevation, and makes adjacent array elements correspond to adjacent physical points in space.

[0050] By copying the first row of data and appending it to the last row in a closed-loop operation, a periodic boundary condition is constructed, which effectively eliminates the discontinuity of the azimuth angle at the periodic boundary and ensures that when calculating the azimuth angle increment of adjacent points in subsequent traversals, all angle intervals, including the connection between the first and last points, can be fully covered.

[0051] S42. Traverse the dataset after loop closure and calculate the azimuth increment of adjacent elements. ,like Then the azimuth increment is corrected to ; S43. Extract the maximum value from all azimuth increments. Second largest value And its corresponding indices m and n, comparing the maximum and second largest values ​​relative to each other. Degree of deviation: like , If the absolute value is taken, the direction of the maximum gap is determined to be the direction of the bridge's external normal. The calculation formula is as follows:

[0052] Otherwise, the direction of the second largest gap is determined to be the direction of the bridge's external normal, and the calculation formula is:

[0053] Normalize the calculated A0. Then it is corrected to ; S44. Calculate the reverse azimuth angle ,like Then it is corrected to .

[0054] S5. Based on the initial edge point Q and the reverse azimuth angle A1, locate the initial edge point W on the opposite side of the bridge.

[0055] Preferably, the positioning process of the initial edge point W on the opposite side of the bridge includes: S51. Traverse the neighborhood dataset P, based on the azimuth angle. Constraints are used to filter the dataset, removing those that do not meet the angle range. The interference point has the following azimuth constraint condition:

[0056] In the formula, α is a preset angle, and its value is negatively correlated with the width of the bridge. That is, the wider the bridge is, the smaller the value of the coefficient α is.

[0057] like Figure 4 As shown, a sector-shaped search region is constructed based on the neighborhood dataset P, and invalid interference points that do not satisfy the angle range α are removed.

[0058] S52. Traverse the dataset after the removal operation in S51, based on elevation... Constraints are used to filter the dataset, removing data that does not meet the elevation threshold. The interference points have the following elevation constraints:

[0059] In the formula, This refers to the critical elevation obtained in step S34.

[0060] S53. Calculate the point with the maximum horizontal distance from each of the remaining points in the dataset to the initial edge point Q, and use it as the initial edge point W on the opposite side of the bridge.

[0061] S6. Using the initial edge point Q and the opposite initial edge point W as starting seed points respectively, based on the azimuth angle A0 and the reverse azimuth angle A1, an iterative stepping strategy is adopted to search for adjacent edge points along both sides of the bridge longitudinal direction. The change in the angle between the current search point and the preceding and following adjacent points is calculated in real time, and the search termination condition is determined by the angle threshold. Finally, a complete array of bridge double-sided edge point sets is generated.

[0062] Preferably, the retrieval process for adjacent edge points is as follows: Figure 5 As shown, the retrieval steps are as follows: S61. Traverse the neighborhood dataset P, initially setting the current node as the initial edge point Q, and so on. , Centered on the initial normal azimuth angle A0, calculate the clockwise angle between the line connecting the current point and Q and A0, and select the point with the minimum angle as the right extension point of A0. Select the point with the maximum included angle as the left extension point of A0. This enables traversal and retrieval of edge points along both sides of the bridge's longitudinal direction. like Figure 6 Demonstration, Point The angle between the line connecting Q and A0 clockwise is the smallest. The angle between the line connecting Q and A0 is the largest. Therefore... This is the right-side extension point, i.e., the next reference point for the roadbed edge line in the right direction; This is the left extension point, i.e., the next reference point for the roadbed edge line in the left direction.

[0063] Each retrieved point undergoes a coordinate transformation to convert its polar coordinate data to coordinates in the construction coordinate system, and is then stored sequentially in arrays AQ1 (right-side point) and AQ2 (right-side point). A

[0064]

[0065] S62. Starting from the initial edge point W on the opposite side of the bridge, repeat the retrieval steps based on the reverse azimuth angle A1 to extract the edge point on the other side of the bridge.

[0066] The retrieved points on the other side of the bridge are then converted to reflect their coordinates and stored in arrays AW1 and AW2.

[0067] S63. When the change in the included angle exceeds a preset included angle threshold, it is determined that the bridge range has been exceeded and the search is stopped.

[0068] Preferably, the included angle threshold γ is preset to be .like Figure 7 The diagram shows the termination of the search. When the included angle dβ of three adjacent edge points is less than γ, it indicates that the edge trajectory has a sharp geometric change (such as an approximately right-angle turn). This usually indicates that the bridge has terminated or reached the edge of the bridge, thus terminating the search.

[0069] After the search is terminated, the arrays AQ1, AQ2, AW1 and AW2 are the bridge edge feature point sets determined by the search algorithm, and their coordinate data constitute the final output of the edge extraction method.

[0070] To verify the effectiveness of the method for automatically extracting bridge edges from point clouds provided in this embodiment, the following experimental data is provided. The original point cloud data of a bridge and its surrounding environment is as follows: Figure 8 As shown, the 2D point cloud view of the bridge is as follows: Figure 9 As shown, the generated edge map is compared and analyzed to verify whether the extracted planar position information is consistent with the preset benchmark. The key preset parameter values ​​involved in the method are as follows: When calculating the initial direction and retrieving edge points, the values ​​of inner diameter R1 and outer diameter R2 are set to 1m and 3m, respectively. When detecting the initial edge point W on the other side of the bridge, α is taken as 0.1 radians; When retrieving edge points, Values .

[0071] By following the steps provided in this embodiment, and using the LiDAR360 software platform for data browsing and processing, the complete double-sided edge lines of the bridge were successfully extracted and output. The extracted bridge edge line results are as follows: Figure 10 As shown.

[0072] Example 2 In a second aspect, the present invention provides a system for automatically extracting bridge edge lines from point clouds, comprising: The data input module is used to load and acquire 3D point cloud data; The initial setup module is used to select initial edge points and set algorithm parameters in the point cloud view; The orientation calculation module is used to perform polar coordinate transformation, elevation change detection, and automatic calculation of the initial normal azimuth angle; The opposite side positioning module is used to determine the starting edge point on the opposite side of the bridge; The iterative detection module is used to execute the edge point retrieval algorithm. The results output module is used to generate and store the bridge edge point array; Preferably, the system can run on an engineering surveying terminal or a cloud server, supports interface docking with a construction management platform, and directly uses the extracted boundary data for spatial positional relationship analysis between new construction projects and existing bridges.

[0073] 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 within the protection scope of the present invention.

Claims

1. A method for automatically extracting bridge edge lines from point clouds, characterized in that, Includes the following steps: S1. Obtain a 3D point cloud dataset containing the bridge; S2. Determine the initial edge point Q in the point cloud dataset; with Q as the center, Inner diameter A ring-shaped neighborhood filtering region is constructed for the outer diameter, filtering all neighborhood points that satisfy the constraints; with Q as the origin of the polar coordinates, the neighborhood points are transformed into polar coordinates and a neighborhood dataset is constructed, represented as: In the formula, P represents the neighborhood dataset. This represents the horizontal distance from point i to Q. This represents the azimuth angle of Q pointing to the i-th point. Indicate the elevation of point i; S3. Detect elevation abrupt changes using a preset method, remove non-bridge points, and obtain a clean bridge point dataset P1; S4. Based on the pure bridge point dataset P1, detect and calculate the initial normal azimuth angle A0 and the reverse azimuth angle A1; S5. Based on the initial edge point Q and the reverse azimuth angle A1, locate the initial edge point W on the opposite side of the bridge; S6. Using the initial edge point Q and the opposite initial edge point W as starting seed points respectively, based on the azimuth angle A0 and the reverse azimuth angle A1, an iterative stepping strategy is adopted to search for adjacent edge points along both sides of the bridge longitudinal direction. The change in the angle between the current search point and the preceding and following adjacent points is calculated in real time, and the search termination condition is determined by the angle threshold. Finally, a complete array of bridge double-sided edge point sets is generated.

2. The method for automatically extracting bridge edge lines from point clouds according to claim 1, characterized in that, In step S1, any point in the 3D point cloud dataset is represented by coordinates in the construction coordinate system. Unique identifier, among which For the north coordinate component, For the eastern coordinate component, This represents the elevation component.

3. The method for automatically extracting bridge edge lines from point clouds according to claim 1, characterized in that, In step S2, the initial edge point Q is located in the edge area of ​​the bridge beam and maintains a preset safe distance from the pier structure. Its spatial coordinates are obtained through on-site measurement or through manual interactive extraction based on the point cloud data.

4. The method for automatically extracting bridge edge lines from point clouds according to claim 1, characterized in that, The constraint conditions mentioned in step S2 are: Where r is the horizontal distance from the neighboring point to the initial edge point Q; inner diameter The value is negatively correlated with the spatial distribution density of the edge points to be extracted; outer diameter The value is less than the horizontal distance from Q to the nearest pier.

5. The method for automatically extracting bridge edge lines from point clouds according to claim 1, characterized in that, The preset method for detecting elevation changes in step S3 is as follows: S31. Based on the stated elevation The neighborhood dataset P is sorted in descending order of size. S32. Traverse and calculate the elevation difference between adjacent data. ; S33. Based on the preset elevation change threshold Judgment, if If the (i+1)th point is an elevation change point, then the (i+1)th element and its subsequent elements are non-bridge points. S34. Record the current critical elevation h0= The dataset is truncated, and the first to the ith elements are retained to obtain the pure bridge point dataset P1. The threshold Set to the minimum height difference between the bottom of the bridge and the ground.

6. The method for automatically extracting bridge edge lines from point clouds according to claim 1, characterized in that, In step S4, the initial normal azimuth angle A0 is perpendicular to the bridge orientation and points outward from the bridge; the reverse azimuth angle A1 is perpendicular to the bridge orientation and points inward from the bridge. The detection calculation process is as follows: S41. Arrange the pure bridge point dataset P1 according to coordinate azimuth angle Sort in ascending order; perform a data loop operation, copying the first element of the sorted dataset and appending it to the end of the dataset to construct periodic boundary conditions; S42. Traverse the dataset after loop closure and calculate the azimuth increment of adjacent elements. ,like Then the azimuth increment is corrected to ; S43. Extract the maximum value from all azimuth increments. Second largest value And its corresponding indices m and n, comparing the maximum and second largest values ​​relative to each other. Degree of deviation: like , If the absolute value is taken, the direction of the maximum gap is determined to be the direction of the bridge's external normal. The calculation formula is as follows: Otherwise, the direction of the second largest gap is determined to be the direction of the bridge's external normal, and the calculation formula is: Normalize the calculated A0. Then it is corrected to ; S44. Calculate the reverse azimuth angle ,like Then it is corrected to .

7. The method for automatically extracting bridge edge lines from point clouds according to claim 5, characterized in that, The positioning process for the initial edge point W on the opposite side of the bridge in step S5 includes: S51. Traverse the neighborhood dataset P, based on the azimuth angle. Constraints are used to filter the dataset, removing those that do not meet the angle range. The interference point has the following azimuth constraint condition: In the formula, α is a preset angle range, and its value is negatively correlated with the width of the bridge; S52. Traverse the dataset after the removal operation in S51, based on elevation... Constraints are used to filter the dataset, removing data that does not meet the elevation threshold. The interference points have the following elevation constraints: In the formula, This is the critical elevation. S53. Calculate the point with the maximum horizontal distance from each of the remaining points in the dataset to the initial edge point Q, and use it as the initial edge point W on the opposite side of the bridge.

8. The method for automatically extracting bridge edge lines from point clouds according to claim 1, characterized in that, The retrieval steps for adjacent edge points in step S6 are as follows: S61. Traverse the neighborhood dataset P, initially setting the current node as the initial edge point Q, and so on. , Centered on the initial normal azimuth angle A0, calculate the clockwise angle between the line connecting the current point and Q and A0, and select the point with the minimum angle as the right extension point of A0. Select the point with the maximum included angle as the left extension point of A0. This enables traversal and retrieval of edge points along both sides of the bridge's longitudinal direction. S62. Taking the initial edge point W on the opposite side of the bridge as the starting point, the retrieval steps are repeated based on the reverse azimuth angle A1 to extract the edge point on the other side of the bridge. S63. When the change in the included angle exceeds a preset included angle threshold, it is determined that the bridge range has been exceeded and the search is stopped.

9. The method for automatically extracting bridge edge lines from point clouds according to claim 1, characterized in that, The included angle threshold γ in step S6 is preset to be .

10. A system for automatically extracting bridge edge lines from point clouds, characterized in that, include: The data input module is used to load and acquire 3D point cloud data; The initial setup module is used to select initial edge points and set algorithm parameters in the point cloud view; The orientation calculation module is used to perform polar coordinate transformation, elevation change detection, and automatic calculation of the initial normal azimuth angle; The opposite side positioning module is used to determine the starting edge point on the opposite side of the bridge; The iterative detection module is used to execute the edge point retrieval algorithm. The results output module is used to generate and store the array of bridge edge points.

Images