Automatic detection method for perpendicularity of large-batch piers based on ground three-dimensional scanning

Abstract

The invention relates to the technical field of pier perpendicularity measurement, and discloses a large-batch pier perpendicularity automatic detection method based on ground three-dimensional scanning, comprising the following steps: S1, arranging a three-dimensional laser scanner and a target ball, and registering single-site cloud collected by the three-dimensional laser scanner based on the target ball to obtain bridge and environment point cloud; the method comprises the steps of S1, extracting bridge point cloud, S3, segmenting the point cloud to obtain a local coordinate system, S4, extracting bridge pier point cloud, and S5, judging the shape of the cross section of the bridge pier; and S6, detecting the perpendicularity of the bridge pier. Therefore, the continuous high-precision bridge pier axis can be fitted by collecting a large amount of point cloud data, and the perpendicularity is calculated through the included angle between the axis equation and the vertical direction by fitting the centroid coordinates of all the point cloud slices of the bridge pier and aiming at different bridge pier section contours, so that the accuracy of the measured data is higher; in addition, the bridge pier point cloud is automatically segmented in a large bridge, and batch detection of the bridge piers is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of pier verticality measurement, and in particular to a method for automatically detecting the verticality of a large number of piers based on terrestrial three-dimensional scanning. Background Art

[0002] At present, small and medium-span bridges account for a relatively high proportion and are the main components of high-speed bridges. As the main beam support structure, the state of the pier will directly affect the safety of the bridge. As one of the important indicators reflecting the state of the pier, it is particularly important to quickly and accurately measure the verticality of the pier during the construction and operation periods.

[0003] The most commonly used methods for detecting the verticality of piers mainly include total stations and bridge inspection vehicles, etc. Traditional detection methods such as total stations are characterized by low efficiency and huge workloads. They can only achieve single-point measurement and need to measure a single pier column multiple times, resulting in a huge workload for verticality detection. In addition, the detection method based on total stations needs to measure in two perpendicular directions of the pier column, which is obviously difficult to achieve in steep mountainous areas and is also difficult to detect the verticality of piers with different cross-sections. Summary of the Invention

[0004] The purpose of the present invention is to overcome the deficiencies of the prior art and provide a method for automatically detecting the verticality of a large number of piers based on terrestrial three-dimensional scanning. This method uses terrestrial laser scanning (TLS) and has the characteristics of fast speed, high precision, and high reliability. For the detection of pier verticality, it can batch-detect the verticality of piers with different cross-sections and in different environments.

[0005] The purpose of the present invention is achieved through the following technical solutions: An automated detection method for the perpendicularity of a large number of bridge piers based on terrestrial three-dimensional scanning, comprising the following steps: S1. Collect single-site point clouds based on a three-dimensional laser scanner and target balls, identify the target ball point clouds within the single-site point clouds, and obtain an overall point cloud model of the bridge and the surrounding environment through the registration of the target ball point clouds; S2. Segment the overall point cloud model to segment out the bridge point clouds; S3. Estimate the bridge route of the bridge, slice the bridge point clouds along the bridge route to obtain a number of first slices, obtain the peak slice among the number of first slices, extract the pier area point clouds from the bridge point clouds based on the peak slice, and establish a local coordinate system for the pier area point clouds; S4. Slice along a second direction in the local coordinate system to obtain a number of second slices, count the point cloud density of each second slice to obtain the first pier point clouds, slice the first pier point clouds along a third direction to obtain third slices, obtain the filtered cross-section size according to the third slices, and obtain the second pier point clouds through bounding box filtering, and segment and extract the third pier point clouds; S5. Discriminate the pier cross-section contour based on the bounding box size and the standard deviation of the fitted circle; S6. For different pier cross-section contours judged in S5, slice the third pier point clouds along a fourth direction in the local coordinate system to obtain fourth slices, calculate the centroid of the pier according to the point clouds of the fourth slices, and fit the centroid of the cross-section to detect the perpendicularity of the pier cross-section.

[0006] The beneficial effects of the present invention are as follows: The method can fit a continuous and high-precision pier axis by collecting a large amount of point cloud data, and adopts the centroid coordinates of all point cloud slices of the fitted pier. For different pier cross-section contours, the perpendicularity is calculated through the angle between the axis equation and the vertical direction, making the measured data more accurate; in addition, in the point cloud containing vegetation, ground, and up to dozens of spans of main girders, cover girders, cross girders, and piers obtained by point cloud registration in this application, a segmentation algorithm for pier point clouds is proposed, realizing the batch and automatic segmentation of pier column point clouds, automatically segmenting pier point clouds in a large-scale bridge, and realizing the batch detection of piers. Brief Description of the Drawings

[0007] Figure 1 It is a schematic flow chart of the automated detection method for the perpendicularity of a large number of bridge piers based on terrestrial three-dimensional scanning according to an embodiment of the present application; Figure 2 It is a schematic flow chart of another expression of the automated detection method for the perpendicularity of a large number of bridge piers based on terrestrial three-dimensional scanning according to an embodiment of the present application. Detailed Embodiments

[0008] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0009] Referring to Figure 1 , the present invention provides a technical solution: A method for automatically detecting the perpendicularity of a large number of bridge piers based on ground three-dimensional scanning according to an embodiment of the present application will be explained for the following professional terms before a detailed description: Single-station point cloud: It refers to the point cloud data obtained from a fixed scanning station (i.e., the position of a laser scanner). During the three-dimensional laser scanning process, the laser scanner scans an object from a specific position and obtains the point cloud data of the object's surface from the perspective of this position, thus forming a single-station point cloud.

[0010] Bounding box: It is a minimum geometric body used to enclose an object or a point cloud, and a bounding box can be generated according to the geometric model of the object.

[0011] Perpendicularity calculation: Calculate the angle or offset by the difference between the slope of the axis equation and the theoretical vertical direction.

[0012] In some embodiments, in combination with Figure 1 - Figure 2 understanding, the method for automatically detecting the perpendicularity of a large number of bridge piers based on ground three-dimensional scanning includes the following steps: S1. The user arranges the three-dimensional laser scanner beside the bridge pier according to the limitation of the incident angle of the ground laser scanner. For example, the scanner is kept at a safe distance of 0.5 - 1.5 meters from the surface of the bridge pier; target balls are arranged on the bridge pier. For example, using a magnetic base for fixation, the target balls are arranged in different areas of the bridge pier.

[0013] Collect a single-station point cloud based on the arranged three-dimensional laser scanner and target balls, and identify the target ball point cloud in the single-station point cloud. The overall point cloud model of the bridge and the surrounding environment is obtained through point cloud registration by the target balls.

[0014] S2. Segment the overall point cloud model to segment out the bridge point cloud, which specifically includes the following steps: S21. Segment the overall point cloud model into above-ground point cloud and ground point cloud using the cloth filtering algorithm.

[0015] It should be noted that the stiffness coefficient d has an important impact on the stiffness of the fabric and the degree of fit between the fabric simulation result and the ground. The fabric filtering algorithm is currently mainly used to process the airborne scanning point cloud of the unmanned aerial vehicle to obtain the ground digital model. The point spacing is generally dozens of centimeters, so the d value adopted is relatively large. In this application, the ground three-dimensional laser scanner is used to collect point cloud data, and the point spacing reaches the millimeter level. Therefore, it is necessary to dynamically adjust the stiffness coefficient d according to the scene to improve the ground point cloud segmentation accuracy. In some preferred examples, d = 10r is selected, where r is the scanner resolution (i.e., the point spacing at a 10m ranging distance). The overall point cloud is segmented into the ground point cloud and the above-ground point cloud.

[0016] S21. Use the Euclidean clustering segmentation algorithm for the above-ground point cloud to segment it into multiple clustering clusters, and segment out the bridge point cloud through the maximum distance between the preset adjacent clustering clusters and the minimum number of point clouds in the clustering cluster.

[0017] In some examples, takes a value of 0.5m, takes a value of 1 / 100 of the total number of point clouds. Furthermore, the above-ground point cloud is segmented into the bridge point cloud and multiple vegetation point cloud clustering clusters. Count the number of point clouds in each clustering cluster after clustering. The first N clustering clusters with the largest number of point clouds are the bridge point cloud, where N is the number of bridge spans.

[0018] S3. Estimate the bridge route of the bridge, slice the bridge point cloud along the bridge route to obtain several first slices, obtain the peak slices among the several first slices, extract the pier area point cloud from the bridge point cloud based on the peak slices, and establish the local coordinate system of the pier area point cloud. The specific steps are as follows: S31. Project the bridge point cloud onto the XOY plane to obtain the two-dimensional bridge point cloud.

[0019] S32. Use principal component analysis to obtain the bridge forward direction, and use a quadratic parabola to fit the centroid coordinates of the point cloud data in the two-dimensional bridge point cloud, where the subscript mi is used to distinguish parameters, so as to obtain the rough bridge route , where a, b, and c are all parameters, x is the horizontal coordinate value variable of the bridge route, and y is the vertical position value variable of the bridge route.

[0020] S33. Slice the projected bridge point cloud (i.e., the two-dimensional bridge point cloud) along the rough bridge route direction at the first slice spacing D to obtain several first slices. In some examples, the value of the first slice spacing D satisfies , is the maximum width of the capping beam along the bridge direction.

[0021] Specifically, the bridge route length I in the area where the bridge point cloud is located is divided into intervals according to the spacing D. I is calculated by the following formula:

[0022] In the formula, and respectively refer to the minimum and maximum values of the centroid coordinates. Furthermore, the coordinates of each equally divided point of the bridge route are obtained, where . Then, the point cloud between adjacent equally divided points is segmented by a straight line passing through the equally divided points. The equation of the straight line passing through the i-th equally divided point is shown as follows.

[0023]

[0024] Count the number of point clouds in the first slice after segmentation, and count the position of the first slice with the largest number of point clouds in the statistical graph, which corresponds to the position of the bridge pier and is later called the peak slice. Use the point clouds in the area on both sides of the tangent direction of the center point route of the peak slice position as the point clouds of the bridge pier area at this position to extract the point clouds of the bridge pier area from the bridge point cloud .

[0025] S35. Determine the point clouds of the bridge pier area based on the local coordinate system , establish the local coordinate system of the bridge pier , take the centroid point in the XOY plane as the coordinate origin , take the straight line perpendicular to the rough bridge route equation passing through the centroid point as the y-axis, and take the tangent line passing through the centroid point as the x-axis. The conversion formula from the local coordinate system of the bridge pier to the global coordinate system XOY is shown as follows:

[0026] In the formula, is the angle between the global coordinate system and the local coordinate system, .

[0027] S4. Obtain a number of second slices by slicing along the second direction in the local coordinate system, count the point cloud density of each second slice to obtain the first bridge pier point cloud, slice the first bridge pier point cloud along the third direction to obtain the third slice, and through bounding box filtering, segment and extract the third bridge pier point cloud; specifically including: S41. Slice along the second direction in the local coordinate system of the point clouds of the bridge pier area. The second direction is the y-axis direction mentioned above, and a number of second slices are obtained. The second slice The value of the spacing is preferably less than 1 / 2 of the minimum size of the pier cross-section, and the purpose is to completely extract the point cloud of the first pier. Thus, second slices are obtained, which are the maximum and minimum values in . If the number of points in each second slice is , then the average number of points in all second slices is . Take

[0028] S43. Slice the point cloud of the first pier along the third direction, where the third direction is the z-axis direction of the local coordinate system, to obtain a number of third slices, and project the point cloud of the third slices onto the XOY plane.

[0029] Fit the edges of the point cloud projected by the third slice with a square frame, and record the size of the square frame corresponding to the point cloud of the third slice , as well as the centroid coordinates corresponding to this square frame . The subscript ri is used to distinguish the parameters, where is the average value of the z values of all points in the third slice, are the centroid coordinates of the square box enclosed in the third slice respectively. Then, statistically analyze the square sizes fitted from all third slices, and take the square size with the highest probability as the filtered cross-section size of the pier column . Due to the influence of the pier verticality, increase the allowable error of the size of the square frame respectively to extract the complete point cloud of the second pier. Then the final filtered square frame size is . In the height direction of the square bounding box, use the fitting line of the centroid corresponding to the square size with the highest probability. After using the bounding box filtering, there is still the point cloud of the main girder inside the square frame where the pier is located. Use clustering segmentation to extract the filtered point cloud of the third pier, and for the sake of convenience of explanation, it is defined as the point cloud of the third pier.

[0030] S5. According to the filtered cross-section size , based on the bounding box size and the standard deviation of the fitted circle, determine the pier cross-section contour and determine the pier cross-section type. Specifically, it includes determining the pier cross-section type according to the following formula: Use the first condition in the formula and to distinguish between rectangular cross-section piers and square or circular cross-section piers. Then, according to the second condition and to distinguish between square cross-section and circular cross-section piers. Specifically, if is satisfied, then the pier is a rectangular cross-section pier; if ​Then the pier is a square or circular pier, and then judge: If it satisfies , then the pier is a square pier. If it satisfies , then the pier is a circular pier. As follows:

[0031]

[0032] Where is the difference between the length and width of the cross-section, used to distinguish between rectangular and square cross-sections, generally taking 0.1 m. is the standard deviation of the circular fitting of the z-value maximum sliced point cloud obtained by slicing the point cloud of the sub-pier along the Z-axis, which is the standard deviation threshold for distinguishing between circular and square cross-sections, and takes a value of 0.1 in some examples.

[0033] In this way, according to the relevant parameters in the pier column point cloud segmentation algorithm, that is, the size of the bounding box, and the fitting standard deviation, the discrimination of circular, square, and rectangular pier cross-sections is realized. Compared with the prior art in which the piers with different cross-sections are not identified, the accuracy of this application is higher.

[0034] S6. For the different pier cross-section contours judged in S5, the fourth slice obtained by slicing the third pier point cloud along the fourth direction in the local coordinate system is used to calculate the centroid of the pier according to the point cloud of the fourth slice, and the perpendicularity of the pier cross-section is detected by fitting the centroid of the cross-section.

[0035] It can be understood that the method of this application can fit a continuous and high-precision pier axis by collecting a large amount of point cloud data, and uses the centroid coordinates of the fitted slices of all the point clouds of the pier. For different pier cross-section contours, the perpendicularity is calculated by the angle between the axis equation and the vertical direction, making the measured data more accurate; in addition, in the point cloud including vegetation, ground, and up to dozens of spans of main girders, cover girders, cross girders, and piers through point cloud registration in this application, a segmentation algorithm for pier point clouds is proposed, realizing batch and automatic segmentation of pier column point clouds, automatically segmenting pier point clouds in large-scale bridges, and realizing batch detection of piers. Compared with the prior art in which batch extraction of pier column point clouds is not carried out, but for a single pier and the target type is the detection method of high piers, the effect is better.

[0036] Next, S6 will be described in detail based on some exemplary embodiments.

[0037] The first embodiment, for a square pier with complete point cloud: S61. Slice the third pier point cloud along the fourth direction. In this example, the fourth direction is the z-axis slice in the local coordinate system to obtain a number of fourth slices. And project the point cloud of the fourth slice onto the XOY plane to obtain a two-dimensional contour point cloud.

[0038] S62. Segment the point cloud of the fourth slice after projection using a square grid with a size of and fit the side line of the pier .

[0039] Specifically, the grid size is , calculated according to the following formula:

[0040] where is the grid size factor, which is in some examples.

[0041] Then, use the minimum number q of the point cloud inside each grid and the standard deviation of the fitted line to remove the unqualified grid point cloud in the side line point cloud. The point cloud inside each grid uses the RANSAC method for line fitting, and records the slope and intercept of the fitted line in each grid, is the intercept on the y-axis. Among them, q can take 5, can take 0.001. Then compare the intercepts of the fitted lines of each grid, and merge the grid internal point clouds into the point cloud of one side line, and use RANSAC to fit the side line equation, so as to obtain each side line equation of the pier cross-section. Among them, is the threshold of the intercept on the y-axis of the fitted line of the grid internal point cloud on the same side line, and its value is .

[0042] S63. Solve the equations of multiple side lines in pairs to obtain several intersection coordinates , where i is a positive integer and j is also a positive integer, represents the i th intersection point in the j th fourth slice, and the two straight lines satisfying the side line equation do not solve for the intersection point. Among them, is the slope threshold to prevent parallel side lines from finding the intersection point, and it can take the value of 2; S64. Calculate the centroid coordinates using the solved intersection coordinates of the side lines. The subscript m is used to distinguish parameters, represents the centroid coordinates i of the th fourth slice. Among them, . The centroid points of the point cloud of each fourth slice are fitted by the least squares fitting method to obtain the pier axis. , where A, B, C, and E are all calculated coefficients, and x, y, and z are the coordinate values of the pier axis corresponding to the coordinate system in three directions; thus, the perpendicularity detection is realized based on the pier axis: S65. Calculate the pier offset angle and the pier top displacement (M is the designed height of the pier):

[0043] , the pier top displacement .

[0044] Second embodiment. For a square pier with missing point cloud, after step S63, S631A is performed.

[0045] S631A: Take the centroid coordinates of the point cloud of the pier cross-section as the initial center coordinates , where the subscript c is used to distinguish parameters, calculate the distances from the initial center coordinates to three side lines, and take the variance between as the objective function, as shown in the following formula:

[0046] Among them, is the mean value of.

[0047] Continuously perform iteration, and then determine the centroid coordinates of the pier column cross-section, where the subscript l is used to distinguish parameters, and the corresponding inscribed circle radius at that time, and , where are the distances from the center coordinates to three sides when the iteration is completed. Take the inscribed circle center coordinates of each fourth slice as the centroid coordinates of the contour of the fourth slice, and fit a straight line to calculate the perpendicularity of the pier.

[0048] It can be understood that for rectangular and square bridge piers with complete point clouds, a calculation method for the cross-sectional contour of the bridge pier is proposed. Four side lines of the bridge pier are obtained, and then the intersection points are calculated pairwise to obtain the corner point coordinates of the bridge pier. Then, the centroid coordinates of the cross-section are calculated based on the average value of the intersection point coordinates between the side lines of the bridge pier cross-section. Finally, the least squares method is used to fit the centroid coordinates to obtain the axis equation, and the verticality of the bridge pier is calculated by calculating the angle between the axis equation and the vertical direction. Thus, based on all the slices of a single bridge pier, the axis equation is obtained by fitting the centroid coordinates, and the verticality is calculated by the angle between the axis equation and the vertical direction, making the result more accurate, and reducing random errors compared to the conventional method of calculating the vertical height using the centroid coordinates at the bottom and top.

[0049] In some embodiments, for a rectangular bridge pier with missing point clouds, after step S63, S631B is performed: First, in the geometric model under ideal conditions, let the intersection points of the two diagonals of the rectangular cross-section be A and B, and use them as the reference to make the circumscribed circle of the rectangle. The radius of this circle is and the center coordinates are . When this circle passes through the minimum coordinate points in the Y-axis direction of the corresponding point cloud data of the first side line and the third side line of the rectangle , two circumscribed circle equations can be established, denoted as the equations of circle 1 and circle 3. If the length of the first side is equal to the length of the third side at this time, that is . Then the center coordinates coincide exactly with the centroid of the rectangular cross-section .

[0050] However, in practice, due to construction errors and the influence of the scanner incident angle, the center coordinates deviate from the centroid. After research, it is found that if the circumscribed circle must pass through A, B, and a certain point C on the third side line (and only C is allowed to move in the Y-axis direction), then the deviation of the center coordinates is only reflected in the y-axis direction. Therefore, is used as the final center coordinate of the circumscribed circle of the rectangular cross-section, which is the centroid coordinate of the actual rectangular cross-section. Finally, the centroid coordinates of the cross-section are extracted layer by layer along the height direction of the bridge pier, and the centroid coordinates of each slice are fitted to obtain the axis equation of the bridge pier, and the verticality of the bridge pier can be calculated through the verticality calculation.

[0051] In some embodiments, for a circular cross-section bridge pier, S6 specifically includes the following steps: S61C: Slice the point cloud of the third bridge pier along the z-axis to obtain the fifth slice, and project the fifth slice onto the XOY plane.

[0052] S62C: Use the three-point circle method to solve the center coordinates and radius for the projected two-dimensional point cloud, and traverse all the projected point clouds of the fifth slice in sequence.

[0053] S63C: Perform probability statistics on the radius to obtain the radius with the highest probability and its corresponding center coordinates. Then, with as the center, construct a toroid with an outer diameter of , an inner diameter of , and a height equal to the thickness of the fifth slice, and delete the point cloud outside the toroid. Among them is the allowable error, the purpose of which is to ensure the integrity of the pier point cloud and can be taken as 5 mm.

[0054] S64C: Fit the centroid coordinates of each fifth slice to obtain the pier axis equation, and thus calculate the pier verticality.

[0055] It can be seen that the embodiments of the present application can calculate the verticality of rectangular and square piers with missing single-sided point clouds based on the methods of inscribed circle iteration and circumscribed circle iteration; and can also calculate the verticality of cylindrical piers, so as to calculate the verticality of square piers and cylindrical piers with complete point clouds and missing single-sided point clouds. Compared with the prior art, it can calculate the verticality of square piers with complete point clouds better. Therefore, compared with the problem that the verticality of piers cannot be calculated when facing piers with missing point clouds. The present application can calculate the verticality of square piers and cylindrical piers with complete point clouds and missing single-sided point clouds, and has a wider applicability.

[0056] The above are only the preferred embodiments of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein, should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be changed within the scope of the concept described herein through the above teachings or the techniques or knowledge in related fields. And the changes and modifications made by those skilled in the art without departing from the spirit and scope of the present invention should fall within the protection scope of the appended claims of the present invention.

Claims

1. An automated detection method for the verticality of a large number of bridge piers based on terrestrial three-dimensional scanning, characterized in that, It includes the following steps: S1. Collect single-site point cloud based on a 3D laser scanner and target balls, identify the target ball point cloud within the single-site point cloud, and obtain the overall point cloud model of the bridge and the surrounding environment through registration of the target ball point cloud; S2. Segment the overall point cloud model to segment out the bridge point cloud; S3. Estimate the bridge route of the bridge, slice the bridge point cloud along the bridge route to obtain a number of first slices, obtain the peak slices among the number of first slices, extract the pier area point cloud from the bridge point cloud based on the peak slices, and establish the local coordinate system of the pier area point cloud; S4. Slice along the second direction in the local coordinate system to obtain a number of second slices, count the point cloud density of each second slice to obtain the first pier point cloud, slice the first pier point cloud along the third direction to obtain the third slice, obtain the filtering section size according to the third slice, and obtain the second pier point cloud through bounding box filtering, and extract the third pier point cloud through Euclidean clustering segmentation; S5. Discriminate the pier cross-section contour based on the bounding box size and the standard deviation of the fitted circle; S6. For different pier cross-section contours judged in S5, slice the third pier point cloud along the fourth direction in the local coordinate system to obtain the fourth slice, calculate the centroid of the pier according to the point cloud of the fourth slice, and fit the centroid of the cross-section to detect the perpendicularity of the pier cross-section.

2. The method for automatically detecting the verticality of a large number of bridge piers based on ground three-dimensional scanning according to claim 1, wherein, The specific content of S2 includes: S21. Segment the overall point cloud model into above-ground point cloud and ground point cloud by using the cloth filtering algorithm; S22. Segment the above-ground point cloud into multiple clusters by using the Euclidean clustering segmentation algorithm, and segment out the bridge point cloud according to the preset parameters.

3. The automated detection method for the verticality of a large number of bridge piers based on ground three-dimensional scanning according to claim 2, characterized in that, The specific content of S3 includes the following steps: S31. Project the bridge point cloud onto the XOY plane to obtain a two-dimensional bridge point cloud; S32. Analyze the forward direction of the bridge and fit the centroid coordinates of the point cloud data in the two-dimensional bridge point cloud , where the subscript mi is used to distinguish parameters, so as to obtain the relational expression describing the bridge route , where a, b, and c are all parameters x is the independent variable coordinate value of the bridge route in the horizontal direction, and y is the position coordinate of the bridge route in the vertical direction; S33. Slice the two-dimensional bridge point cloud along the bridge route at a set first slicing interval D to obtain a number of the first slices; in the direction of; S34. Count the number of point clouds of the first slice after segmentation. The position of the first slice with the largest number of point clouds is the peak slice. Use the point clouds within the area on the left and right of the tangent direction of the central point route at the peak slice position as the pier area point clouds where is the maximum width of the cross-bridge bent cap , where is the maximum width of the cross-bridge bent cap S35. Establish the local coordinate system of the bridge pier corresponding to the two-dimensional bridge point cloud , where the subscript qi is used to distinguish parameters, and the centroid point coordinates in the XOY plane are used as the coordinate origin, and the centroid point coordinates and the straight line perpendicular to the bridge route are used as the y-axis, and the tangent line passing through the centroid point coordinates is used as the x-axis.

4. The method for automatically detecting the verticality of a large number of bridge piers based on terrestrial 3D scanning according to claim 3, characterized in that, The specific content of S4 includes the following steps: S42. Obtain a plurality of the second slices by slicing along the y-axis direction in the local coordinate system, and obtain the number of point clouds in each of the second slices and the average number of point clouds of all the second slices , take and merge the point clouds of it and its adjacent second slices into a single point cloud of the first pier; S43. Slice the first pier point cloud along the Z-axis direction of the local coordinate system to obtain a number of the third slices, and project the point cloud of the third slices onto the XOY plane; fit the edges of the point cloud projected by the third slices with a square frame, and record the size of the square frame corresponding to the point cloud of the third slices as , and the centroid coordinates corresponding to the square frame , with the subscript ri used to distinguish the parameters; statistically analyze the square sizes of the square frames fitted from all the third slices, and take the square size with the highest probability as the filtered cross-section size of the pier column as , where the height direction of the bounding box adopts the fitting straight line of the centroid corresponding to the square size with the highest probability; S44. After using the bounding box filtering, use clustering segmentation to extract the third pier point cloud.

5. The method for automatically detecting the perpendicularity of a large number of bridge piers based on ground three-dimensional scanning according to claim 4, wherein, In S43, the dimensions of the square frame are respectively increased by the allowable error to extract the complete point cloud of the second pier, and the final filtered square frame size is .

6. The method for automatically detecting the verticality of a large number of bridge piers based on ground three-dimensional scanning according to claim 5, wherein The specific content of S5 includes: If the following conditions are met then the pier is a pier with a rectangular cross-section. If the following conditions are met then the pier is a square or circular pier, and then make the following judgment: If is satisfied, then the pier is a square pier. If is satisfied, then the pier is a circular pier; Among them, is the difference between the length and width of the cross-section, is the standard deviation of the circle fitting of the point cloud of the z-value maximum slice obtained by slicing the point cloud of the third pier along the Z-axis, is the standard deviation threshold for distinguishing between circular cross-sections and square cross-sections.

7. The method for automatically detecting the perpendicularity of a large number of bridge piers based on ground three-dimensional scanning according to claim 6, wherein: The specific content of S6 includes: S61. Slice the third pier point cloud along the z-axis in the local coordinate system to obtain a number of the fourth slices, and project the point cloud of the fourth slices onto the XOY plane to obtain a two-dimensional contour point cloud; S62. Divide the point cloud of the projected fourth slice using a square grid, and fit the equations of multiple side lines of the bridge pier. ; S63. Solve the equations of multiple pairs of the side lines to obtain several intersection coordinates , where both i and j are positive integers, represents the i th intersection point in the j th fourth slice; S64. Adopt the coordinates of each of the obtained intersection points Calculate the centroid coordinates , and adopt the least squares fitting method to fit the centroid coordinates of the point cloud of each of the fourth slices , so as to obtain the relational expression of the pier axis , where A, B, C, and E are all parameters, and x, y, and z are the coordinate values of the coordinate system corresponding to the pier axis in three directions; S65. Calculate the offset angle of the pier and the displacement at the top of the pier .

8. The method for automatically detecting the verticality of a large number of bridge piers based on ground three-dimensional scanning according to claim 7, wherein: For a square pier with missing point cloud, perform S631A after step S63; Taking the centroid coordinates of the point cloud of the pier cross-section as the initial center coordinates, where the subscript c is used to distinguish parameters, calculate the distances from the initial center coordinates to any three side lines of the pier cross-section , and use the variance between the three side lines as the objective function, and iterate the objective function to determine the centroid coordinates of the pier cross-section , where the subscript l is used to distinguish parameters, and the inscribed circle radius corresponding to the pier cross-section , and , in the formula are the distances from the center coordinates to the three sides when the iteration is completed respectively; take the inscribed circle center coordinates of each of the fourth slices as the centroid coordinates of the contour of the fourth slice.

9. The method for automatically detecting the verticality of a large number of bridge piers based on ground three-dimensional scanning according to claim 7, wherein For a rectangular pier with missing point cloud, perform S631B after step S63; S631B includes: taking the intersection point of the two diagonals of the rectangular cross-section of the rectangular cross-section pier as a reference to make a circumcircle of the rectangle, obtaining the equations of the corresponding two circumcircles, and the radii of the two circumcircles are r l1 , r l3 , and the center coordinates are respectively; when , the center coordinate coincides with the centroid of the rectangular cross-section; when , based on the center coordinate at , take as the circumcircle center coordinate of the rectangular cross-section, and this circumcircle center coordinate is the actual centroid coordinate of the rectangular cross-section.

10. The automated detection method for the verticality of a large number of bridge piers based on terrestrial three-dimensional scanning according to claim 6, characterized in that: For a pier with a circular cross-section, S6 specifically includes the following steps: S61C. Slice the pier point cloud along the z-axis to obtain the fifth slice, and project the slice onto the XOY plane; S62C. Use the three-point circle method to solve the center coordinates and radius of the circle for the projected two-dimensional point cloud, and traverse all the projected point clouds of the fifth slices in sequence; S63C. Perform probability statistics on the radius with respect to the center of the circle to obtain the radius with the highest probability and its corresponding center coordinates , and then, with as the center, construct a toroid with an outer diameter of , an inner diameter of , and a height equal to the thickness of the fifth slice, and delete the point cloud outside the toroid; where is the allowable error; S64C. Fit the centroid coordinates of each fifth slice to obtain the pier axis equation, thereby calculating the pier perpendicularity.

Images