Suspender cable main cable clamp slip detection method based on three-dimensional laser scanning

Abstract

The application discloses a kind of based on three-dimensional laser scanning's suspension bridge main cable clamp slip detection method, it is related to suspension bridge cable clamp slip detection technical field, comprising the following steps: obtaining main cable full section original point cloud data, and carry out point cloud splicing and cutting;To the data after splicing and cutting, eliminate and point cloud smoothing processing are carried out to redundant point cloud information;To the point cloud data after pre-processing, coloring, removing isolated points and triangular mesh optimization processing are handled, and as first scanning point cloud data model;According to the inspection frequency regulation of main cable inspection in bridge maintenance manual, repeat the above steps, obtain second scanning point cloud data model;According to the registration method of ICP, first scanning point cloud data model and second scanning point cloud data model are registered to point cloud data, and obtain suspension bridge main cable clamp slip detection result.The application solves the problems of low detection efficiency, poor accuracy and strong subjectivity in traditional detection method.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of suspension bridge cable clamp slip detection, and particularly relates to a suspension bridge main cable clamp slip detection method based on three-dimensional laser scanning. BACKGROUND

[0002] As a special bridge structure, the main cable clamp slip of a suspension bridge is one of the important factors causing the instability of the bridge. The cable clamp, as an important structure of the cable system of the suspension bridge, transmits the hanger cable tension generated by the dead load and the live load of the vehicle to the main cable. By applying a pre-tightening force to the screw rod of the cable clamp, the radial component of the hanger cable tension is resisted, thereby avoiding the slip of the cable clamp. If the pre-tightening force of the cable clamp screw rod is too small, the cable clamp will slip, causing the redistribution of the tension of the adjacent hanger cable, and further changing the line shape of the main beam. This process has a divergent effect, and the result is irreversible, which is extremely difficult to recover. At the same time, the slip of the cable clamp will cause the damage of the cable clamp ring joint (especially the upper end ring joint), so that rainwater enters the inside of the main cable, causing the corrosion of the main cable wire, and seriously affecting the durability of the structure.

[0003] The traditional suspension bridge main cable clamp slip detection method mainly includes visual inspection, manual detection, friction force test and special instrument measurement. However, the traditional detection method has the problems of low detection efficiency, poor accuracy and strong subjectivity. SUMMARY

[0004] In view of the above problems in the prior art, the suspension bridge main cable clamp slip detection method based on three-dimensional laser scanning provided by the present application solves the problems of low detection efficiency, poor accuracy and strong subjectivity of the traditional detection method.

[0005] In order to achieve the above-mentioned purposes, the technical scheme adopted by the present application is as follows: a suspension bridge main cable clamp slip detection method based on three-dimensional laser scanning, characterized by comprising the following steps:

[0006] S1: obtaining the main cable full-section original point cloud data by using the main cable unmanned inspection vehicle, and performing point cloud splicing and cutting on the main cable full-section original point cloud data;

[0007] S2: eliminating the redundant point cloud information and performing point cloud smoothing processing on the spliced and cut data, so as to realize data preprocessing;

[0008] S3: performing coloring, removing isolated points and triangular mesh optimization processing on the preprocessed point cloud data, so as to obtain a polygon point cloud model with optimized surface quality, and the polygon point cloud model is taken as a first-stage scanning point cloud data model;

[0009] S4: according to the inspection frequency regulation of the main cable inspection in the bridge maintenance manual, repeating the above steps S1-S3 to obtain a second-stage scanning point cloud data model;

[0010] S5: According to the registration method of ICP, the point cloud data registration is performed on the first period scanning point cloud data model and the second period scanning point cloud data model, and a main cable clamp slip detection result of the suspension bridge is obtained.

[0011] The above-mentioned scheme has the beneficial effects that the suspension bridge cable clamp slip is detected by using the three-dimensional laser scanning method, the original point cloud data is obtained by using the main cable unmanned inspection vehicle, the two period point cloud data is obtained by data processing, the two period point cloud data is registered, and the cable clamp slip detection result is obtained. The present application solves the problems of low detection efficiency, poor accuracy and strong subjectivity of the traditional detection method.

[0012] Further, S1 includes the following sub-steps:

[0013] S1-1: The main cable unmanned inspection vehicle assembly is carried to the lowest point of the main cable in the span, and the unmanned inspection vehicle frame, instrument support, three-dimensional laser scanner and CCD camera are installed;

[0014] S1-2: Based on the installed unmanned inspection vehicle frame, instrument support, three-dimensional laser scanner and CCD camera, the driving system, control system and motion function of the unmanned inspection vehicle are checked, and the data acquisition system and data transmission system of the three-dimensional laser scanner are tested;

[0015] S1-3: After testing, the main cable unmanned inspection vehicle is remotely controlled to run from the span to the main tower direction, and the points are sampled along the radial direction of the main cable, and the original point cloud data of the full section of the main cable is obtained;

[0016] S1-4: The point cloud coordinate normalization processing is performed on the original point cloud data of the full section of the main cable;

[0017] S1-5: The related software is used to select the automatic splicing based on the target, and the point cloud splicing and cutting of the data after the point cloud coordinate normalization processing are performed.

[0018] The above-mentioned further scheme has the beneficial effects that by the above-mentioned technical scheme, the original point cloud data is collected by running the installed unmanned inspection vehicle and each device, and the point cloud data is normalized, spliced and cut.

[0019] Further, the normalization in S1-4 includes the following formula:

[0020]

[0021]

[0022]

[0023]

[0024] wherein (x, y, z) is the normalized point cloud coordinate, (a, b, g) is the rotation parameter, (x0, y0, z0) is the point p to be normalized in the instrument coordinate system, (Ax, Ay, Az) is the translation parameter, R(a), R(b) and R(g) are the rotation matrices of the point p in the new three-dimensional coordinate system around the x-axis, y-axis and z-axis.

[0025] The above further scheme has the beneficial effect that the normalization processing of the point cloud data is realized by the above formula, and the point cloud coordinate of the new coordinate system can be obtained by using the formula, thereby providing good data files for subsequent point cloud registration.

[0026] Further, the elimination of redundant point cloud information in S2 is to select points that meet the modeling requirements and can reduce the amount of calculation according to the amount of point cloud data and data requirements, to determine the minimum neighborhood and initial radius, to perform multiple iterations, and to calculate the point cloud distance by using the Euler distance formula, which is

[0027]

[0028] maxd = max {dist(i, j) | 0≤i≤n, 0≤j≤E}

[0029] mind = min {dist(i, j) | 0≤i≤n, 0≤j≤E}

[0030] dist = maxd - mind

[0031] wherein d(i, j) is the distance between point cloud data p i and p j , (x i , y i , z i ) is the coordinate of point cloud data p i , (x j , y j , z j ) is the coordinate of point cloud data p j , maxd is the maximum value of the distance between two point clouds, mind is the minimum value of the distance between two point clouds, max is the maximum value, min is the minimum value, i and j are two different point clouds, n is the number of point clouds, E is the point cloud set, and dist is the difference between the maximum value and the minimum value of the distance between two point clouds.

[0032] The adjacent point a N of the point cloud value a i and the random point a N1 of the adjacent region beyond the minimum neighborhood number is

[0033] a N = {dist(i, j) < esp | 0≤j≤E}

[0034] a N1 ={a N1 |0≤j≤E}

[0035] wherein, esp represents neighborhood.

[0036] The beneficial effect of the further scheme is that the points meeting the modeling requirements and reducing the calculation amount are selected according to the point cloud distance, and the points meeting the requirements are eliminated, so that the accurate scanning point cloud data model is established.

[0037] Further, the initial radius determination method is that the distance between the random two point clouds is evenly divided into e sections, the frequency a k of d(i,j) in each section is detected, the median of the section where the highest frequency is located is the initial radius, and the highest frequency rang formula is

[0038] rang=max{a k |0≤k≤e}

[0039] wherein, k is the number of the points closest to the distance query point.

[0040] The beneficial effect of the further scheme is that the distance is processed by segmentation according to the frequency value of each section, and the median of the section with the highest frequency is taken as the initial radius.

[0041] Further, the point cloud smoothing processing in S2 adopts a bilateral filtering method.

[0042] The beneficial effect of the further scheme is that the point cloud smoothing processing is performed on the data after the elimination of the redundant point cloud information, and the bilateral filtering method can be used to achieve the point cloud smoothing processing.

[0043] Further, the triangular mesh optimization processing in S3 includes constructing a triangular mesh model of the point cloud data, deleting the nail-shaped objects in the triangular mesh model, and correcting the wrong polygon mesh to obtain a polygon point cloud model with optimized surface quality.

[0044] The beneficial effect of the further scheme is that after the splicing, cutting and preprocessing of the point cloud are completed, the point cloud data is processed by using related software, and since there is no obvious topological relationship between the points of part of the point cloud data, the triangular mesh optimization processing is required, the point cloud data is constructed into a triangular mesh model by using the "encapsulation" command, the nail-shaped objects in the model are deleted, and the wrong polygon mesh is corrected to obtain a polygon point cloud model with optimized surface quality.

[0045] Further, S5 includes the following steps:

[0046] S5-1: based on the ICP registration method, the first scanning point cloud data model is taken as the source point cloud Ps The second period scanning point cloud data model is taken as the target point cloud P t ;

[0047] S5-2: The source point cloud P s and the target point cloud P t are registered and aligned in the three-dimensional space, and the formula is:

[0048] P t = RP s + T

[0049]

[0050]

[0051] r 11 = cos y cos b

[0052] r 12 = -sin y cos a + cos y sin b sin a

[0053] r 13 = sin y sin a + cos y sin b cos a

[0054] r 21 = sin y cos b

[0055] r 22 = cos y cos a + sin y sin b sin a

[0056] r 23 = -cos y sin a + sin y sin b cos a

[0057] r 31 = -sin b

[0058] r 32 = cos b sin a

[0059] r 33 = cos b cos a

[0060] Wherein, R is a rotation matrix, T is a translation matrix, r 11 , r 12 , r 13 , r 21 , r 22 , r 23 , r 31 , r 32 and r 33 are elements of the rotation matrix R, t x , t y and t z are elements of the translation matrix T.

[0061] S5-3: Visualizing the two-period point cloud data aligned by registration to obtain the main cable clamp slip detection result of the suspension bridge.

[0062] The beneficial effect of the further scheme is that the registration of the two-period point cloud data is realized by the above technical scheme, the two-period target point clouds are aligned in space, and the clamp slip detection result is obtained through visual processing. BRIEF DESCRIPTION OF DRAWINGS

[0063] Figure 1 It is a flow chart of a main cable clamp slip detection method of a suspension bridge based on three-dimensional laser scanning.

[0064] Figure 2 It is a structure diagram of a three-dimensional laser scanning device.

[0065] 1, three-dimensional laser scanner; 2, main cable and clamp; 3, CCD camera; 4, unmanned inspection vehicle.

[0066] Figure 3 It is a point cloud data schematic diagram.

[0067] Figure 4 It is a point cloud data preprocessing schematic diagram.

[0068] Figure 5 It is a two-period point cloud registration schematic diagram. DETAILED DESCRIPTION

[0069] The application will be further described below in combination with the drawings and specific embodiments.

[0070] As shown in the drawings, a main cable clamp slip detection method of a suspension bridge based on three-dimensional laser scanning comprises the following steps: Figure 1

[0071] S1: Obtain the main cable full-section original point cloud data by using the main cable unmanned inspection vehicle, and perform point cloud splicing and cutting on the main cable full-section original point cloud data;

[0072] S2: Perform redundant point cloud information elimination and point cloud smoothing processing on the spliced and cut data to realize data preprocessing;

[0073] S3: Perform coloring, outlier removal, and triangular mesh optimization processing on the preprocessed point cloud data to obtain a polygon point cloud model with optimized surface quality, and take the polygon point cloud model as a first-period scanning point cloud data model;

[0074] S4: According to the inspection frequency regulation of the main cable inspection in the bridge maintenance manual, repeat the above steps S1-S3 to obtain a second-period scanning point cloud data model, and the first-period scanning point cloud data model and the second-period scanning point cloud data model are different in inspection frequency; ​

[0075] S5: According to the registration method of ICP, the point cloud data registration is performed on the first period scanning point cloud data model and the second period scanning point cloud data model, and a main cable clamp slip detection result of the suspension bridge is obtained.

[0076] S1 includes the following sub-steps:

[0077] S1-1: The main cable unmanned inspection vehicle assembly is carried to the lowest point of the main cable in the middle of the span, and the unmanned inspection vehicle frame, the instrument support, the three-dimensional laser scanner and the CCD camera are installed.

[0078] In one embodiment of the present application, as shown in Figure 2 The CCD camera is fixedly installed on the unmanned inspection vehicle and faces the detection surface. The slip amount of the cable clamp of the suspension bridge is detected through the movement of the trolley. The three-dimensional laser scanner is fixedly installed on the ring support of the unmanned inspection vehicle and faces the main cable and the cable clamp. The scanning surface is set to have a certain overlap range to facilitate point cloud splicing. The main cable and each cable clamp of the suspension bridge are scanned in the radial full-length range through the movement of the unmanned inspection vehicle. The slip amount is detected through the comparison of the two period scanning point cloud data.

[0079] S1-2: Based on the installed unmanned inspection vehicle frame, instrument support, three-dimensional laser scanner and CCD camera, the driving system, control system and motion function of the unmanned inspection vehicle are checked, and the data acquisition system and data transmission system of the three-dimensional laser scanner are tested.

[0080] S1-3: After testing, the main cable unmanned inspection vehicle is remotely controlled to run from the middle of the span to the main tower direction, and the points are sampled along the radial direction of the main cable by 360°, and the original point cloud data of the full section of the main cable are obtained, as shown in Figure 3

[0081] S1-4: The point cloud coordinate normalization processing is performed on the original point cloud data of the full section of the main cable.

[0082] S1-5: The related software is used, and the cloud compare software is selected in the embodiment. The point cloud splicing and cutting are performed on the data after the point cloud coordinate normalization processing by selecting the automatic splicing based on the target.

[0083] The normalization in S1-4 includes the following formula:

[0084]

[0085]

[0086]

[0087]

[0088] ​Where (x,y,z) are the normalized point cloud coordinates, (α,β,γ) are rotation parameters, (x0,y0,z0) is the point p to be normalized in the instrument coordinate system, (Δx,Δy,Δz) are translation parameters, and R(α), R(β) and R(γ) are the rotation matrices of point p around the x-axis, y-axis and z-axis in the new three-dimensional coordinate system.

[0089] like Figure 4 As shown, in S2, redundant point cloud information elimination involves selecting points that meet modeling requirements and reduce computational cost based on the amount of point cloud data and data needs. This is achieved by determining the minimum neighborhood and initial radius, performing multiple iterations, and then calculating the point cloud distance using the Euler distance formula. The formula is as follows:

[0090]

[0091] maxd=max{dist(i,j)|0≤i≤n,0≤j≤E}

[0092] mind=min{dist(i,j)|0≤i≤n,0≤j≤E}

[0093] dist = maxd - mind

[0094] Where d(i,j) is the point cloud data p i and p j The distance, (x) i ,y i ,z i (p) represents point cloud data. i The coordinates, (x j ,y j ,z j (p) represents point cloud data. j The coordinates are: maxd is the maximum distance between two point clouds, mind is the minimum distance between two point clouds, max is the maximum value, min is the minimum value, i and j are two different point clouds, n is the number of point clouds, E is the point cloud set, and dist is the difference between the maximum and minimum distances between two point clouds.

[0095] Point cloud value a that exceeds the minimum number of neighboring regions N and random point a i Adjacent point a N1 They are respectively

[0096] a N ={dist(i,j) <esp|0≤j≤E}

[0097] a N1 ={a N1 ≥esp|0≤j≤E}

[0098] wherein, esp represents a neighborhood.

[0099] The initial radius determination method is to divide the distance between the random two point clouds into e segments, and detect the frequency a of d(i,j) in each segment k The median of the segment where the highest frequency is located is the initial radius, and the highest frequency rang formula is

[0100] rang=max{a k |0≤k≤e}

[0101] wherein, k is the number of points closest to the query point.

[0102] The point cloud smoothing processing in S2 adopts a bilateral filtering method.

[0103] The triangular mesh optimization processing in S3 includes constructing a triangular mesh model of point cloud data, deleting nail-shaped objects in the triangular mesh model, and correcting the wrong polygon mesh to obtain a polygon point cloud model with optimized surface quality.

[0104] S5 includes the following steps:

[0105] S5-1: Based on the ICP registration method, the first scanning point cloud data model is taken as the source point cloud P s , and the second scanning point cloud data model is taken as the target point cloud P t .

[0106] S5-2: As shown in Figure 5 , the source point cloud P s and the target point cloud P t are registered and aligned in three-dimensional space, and the formula is:

[0107] P t =RP s +T

[0108]

[0109]

[0110] r 11 =cosγcosβ

[0111] r 12 =-sinγcosα+cosγsinβsinα

[0112] r 13 =sinγsinα+cosγsinβcosα

[0113] r 21 =sinγcosβ

[0114] r22 = cos y cos alpha + sin y sin beta sin alpha

[0115] r 23 = -cos y sin alpha + sin y sin beta cos alpha

[0116] r 31 = -sin beta

[0117] r 32 = cos beta sin alpha

[0118] r 33 = cos beta cos alpha

[0119] wherein R is a rotation matrix, T is a translation matrix, r 11 , r 12 , r 13 , r 21 , r 22 , r 23 , r 31 , r 32 and r 33 are elements of the rotation matrix R, t x , t y and t z are elements of the translation matrix T;

[0120] S5-3: Visualizing the two-phase point cloud data aligned by registration to obtain the main cable clamp slip detection result of the suspension bridge.

[0121] The device part of the present application comprises a laser scanning device, a data processing unit and a control system, and the main cable clamp of the suspension bridge is accurately and efficiently detected by the three-dimensional laser scanning technology. The method has the advantages of high detection accuracy, simple operation, strong real-time performance and the like, and can effectively prevent the safety hazards caused by the slip of the main cable clamp of the suspension bridge.

[0122] Those skilled in the art will appreciate that the embodiments described herein are presented for the purpose of helping the reader to understand the principles of the present application, and should be understood as not limiting the scope of protection of the present application to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations according to the technical inspirations disclosed in the present application without departing from the essence of the present application, and these modifications and combinations are still within the scope of protection of the present application.

Claims

1. A method for detecting slippage of main cable clamps in suspension bridges based on three-dimensional laser scanning, characterized in that, Includes the following steps: S1: Use the unmanned inspection vehicle for the main cable to obtain the original point cloud data of the entire cross section of the main cable, and then perform point cloud stitching and cropping on the original point cloud data of the entire cross section of the main cable. S2: Perform redundant point cloud information elimination and point cloud smoothing on the spliced ​​and cropped data to achieve data preprocessing; S3: Color the preprocessed point cloud data, remove isolated points, and optimize the triangular mesh to obtain a polygonal point cloud model with optimized surface quality, which will be used as the first phase of scanning point cloud data model. S4: According to the inspection frequency regulations for main cable inspection in the bridge maintenance manual, repeat the above steps S1-S3 to obtain the second phase of scanned point cloud data model. S5: Based on the ICP registration method, perform point cloud data registration on the first phase scan point cloud data model and the second phase scan point cloud data model to obtain the main cable clamp slip detection results of the suspension bridge. S5 includes the following sub-steps: S5-1: ICP-based registration method, using the first phase scanned point cloud data model as the source point cloud. The second phase of scanned point cloud data model will be used as the target point cloud. ; S5-2: Source Point Cloud and target point cloud For registration and alignment in three-dimensional space, the formula is: in, For rotation matrix, It is a translation matrix. , , , , , , , and All are rotation matrices elements, , and All are translation matrices elements, These are rotation parameters; S5-3: Visualize the registered and aligned point cloud data from the two phases to obtain the slippage detection results of the main cable clamps of the suspension bridge.

2. The method for detecting slippage of main cable clamps in suspension bridges based on three-dimensional laser scanning according to claim 1, characterized in that, S1 includes the following sub-steps: S1-1: Transport the prefabricated components of the unmanned inspection vehicle for the main cable to the lowest point of the main cable mid-span, and install the unmanned inspection vehicle frame, instrument bracket, 3D laser scanner and CCD camera; S1-2: Based on the installed unmanned inspection vehicle frame, instrument bracket, 3D laser scanner and CCD camera, inspect the unmanned inspection vehicle's drive system, control system and motion function, and test the 3D laser scanner's data acquisition system and data transmission system; S1-3: After the test is normal, the remote-controlled unmanned inspection vehicle for the main cable runs from the middle of the span towards the main tower and collects points along the radial direction of the main cable 360° to obtain the original point cloud data of the entire cross section of the main cable. S1-4: Perform point cloud coordinate normalization processing on the original point cloud data of the entire cross section of the main cable; S1-5: Using relevant software, select target-based automatic stitching to stitch and crop the point cloud data after the point cloud coordinates have been normalized.

3. The method for detecting slippage of main cable clamps in suspension bridges based on three-dimensional laser scanning according to claim 2, characterized in that, The normalization in S1-4 includes the following formula: in, The coordinates of the point cloud are normalized. For rotation parameters, Points to be normalized in the instrument coordinate system , For translation parameters, , and For point orbiting in the new three-dimensional coordinate system axis, shaft and The rotation matrix of the axis.

4. The method for detecting slippage of main cable clamps in suspension bridges based on three-dimensional laser scanning according to claim 3, characterized in that, In step S2, redundant point cloud information elimination involves selecting points that meet modeling requirements and reduce computational cost based on the amount of point cloud data and data needs. This is achieved by determining the minimum neighborhood and initial radius, performing multiple iterations, and calculating the point cloud distance using the Euler distance formula. The formula is as follows: in, For point cloud data and distance, For point cloud data coordinates For point cloud data coordinates The maximum value of the distance between two point clouds. The minimum distance between two point clouds. To obtain the maximum value, To obtain the minimum value, and For two different point clouds, For the number of point clouds, For point cloud collection, This is the difference between the maximum and minimum distances between two point clouds; Point cloud values ​​exceeding the minimum number of neighboring regions and random points adjacent points They are respectively in, Indicates the neighborhood.

5. The method for detecting slippage of main cable clamps in suspension bridges based on three-dimensional laser scanning according to claim 4, characterized in that, The method for determining the initial radius is as follows: the distance between two random point clouds is divided into equal parts. Segment, detect within each segment frequency The median of the segment containing the highest frequency number is the initial radius, and the highest frequency number is... The formula is in, This represents the number of points closest to the query point.

6. The method for detecting slippage of main cable clamps in suspension bridges based on three-dimensional laser scanning according to claim 5, characterized in that, The point cloud smoothing process in S2 uses a bilateral filtering method.

7. The method for detecting slippage of main cable clamps in suspension bridges based on three-dimensional laser scanning according to claim 6, characterized in that, The triangular mesh optimization process in S3 includes constructing a triangular mesh model of point cloud data, deleting spikes in the triangular mesh model, and correcting erroneous polygonal meshes to obtain a polygonal point cloud model with optimized surface quality.

Images