Bridge support inspection method based on unmanned aerial vehicle

Abstract

The invention relates to the technical field of unmanned aerial vehicles and bridge inspection, and discloses a bridge support inspection method based on an unmanned aerial vehicle. The method comprises the following steps: importing a bridge three-dimensional point cloud model and a reference slice elevation, and extracting a bridge pier contour; calculating an acquisition interval according to the load parameters, and determining a bridge pier center line as an inspection baseline; buffer inspection waypoints are generated on the two sides of the center line, and the nearest waypoint is selected as an inspection starting point; generating an environment map comprising the bridge piers and the surrounding preset range of the bridge piers; and planning a path by adopting an A-Star algorithm, and carrying out smooth processing through a B spline algorithm to generate an inspection path. Various technical means are fused, an efficient, safe and accurate bridge inspection solution is provided, the manual operation risk and cost are reduced, the inspection efficiency is improved, the flight stability is optimized, and technical support is provided for future intelligent bridge inspection.

Description

Technical Field

[0001] The present invention relates to the technical fields of unmanned aerial vehicles and bridge inspection, and particularly relates to a method for inspecting bridge bearings based on unmanned aerial vehicles. Background Art

[0002] With the acceleration of the urbanization process, bridges, as an important part of transportation infrastructure, their structural safety and reliability are crucial for transportation. Traditional bridge inspection methods usually rely on manual labor or fixed equipment, with low efficiency and difficulty in ensuring inspection accuracy. Especially for the inspection of the bearings between bridge piers and beams, manual inspection has certain risks and limitations. Therefore, using unmanned aerial vehicles for bridge inspection, especially for the inspection of bridge pier bearings, has important research and application value. Summary of the Invention

[0003] Aiming at the above-mentioned existing technical deficiencies, the purpose of the present invention is to provide a method for inspecting bridge bearings based on unmanned aerial vehicles, which solves the problems of low inspection efficiency, insufficient accuracy, and high risk of manual inspection in the prior art.

[0004] To solve the above technical problems, the present invention adopts the following technical solutions: In the first aspect, the present invention provides a method for inspecting bridge bearings based on unmanned aerial vehicles, and the method includes: Step S100: Import the three-dimensional point cloud model of the bridge and the reference slice elevation, perform slicing processing on the point cloud model based on the reference slice elevation, and extract the bridge pier contour; Step S200: Calculate the acquisition interval in the horizontal direction according to the load parameters, where the load parameters include the relative flight distance, camera focal length, sensor size, and resolution, determine the center line of the bridge pier based on principal component analysis, and use this center line as the inspection baseline; Step S300: Generate buffer inspection waypoints on both sides of the center line according to the acquisition interval in the horizontal direction, calculate the distances between the take-off point of the unmanned aerial vehicle and each waypoint, and select the waypoint with the shortest distance as the starting point of the inspection; Step S400: Generate an environmental map including the bridge pier and its surrounding preset range, where the environmental map includes information on the internal structure of the bridge and external obstacles; Step S500: Based on the environmental map, use the A-Star algorithm for path planning and smooth the path through the B-spline algorithm to generate an inspection path.

[0005] Preferably, in a possible implementation manner of the first aspect, the extraction of the bridge pier contour in step S100 includes: Perform thinning processing on the three-dimensional point cloud data, and separate the bridge pier point set based on the DBSCAN clustering algorithm; Determine the main direction of the pier through principal component analysis, calculate the normal vector of each point in the point cloud, and filter out the points parallel or perpendicular to the pier surface; Extract the minimum rotated rectangle of the filtered point set, and obtain the coordinates of the four corner points of the rectangle as the pier contour coordinates.

[0006] Preferably, in a possible implementation manner of the first aspect, the determination of the pier center line in step S200 includes: Based on the principal component analysis algorithm, project the pier point cloud data onto the most representative direction to determine the main direction of the pier; Calculate the center line of the pier according to the main direction, and use the center line as the inspection baseline.

[0007] Preferably, in a possible implementation manner of the first aspect, the generation of the inspection waypoints in step S300 includes: Generate buffer areas on both sides of the center line according to the acquisition interval in the horizontal direction; Generate inspection waypoints based on the buffer areas, calculate the distances between each waypoint and the UAV take-off point, and select the waypoint with the shortest distance as the starting point of the inspection.

[0008] Preferably, in a possible implementation manner of the first aspect, the generation of the environmental map in step S400 includes: Generate an environmental map including the pier and its surrounding preset range based on the three-dimensional point cloud data; Mark the areas outside the preset range as obstacles, and combine the internal structure of the bridge and external obstacle information to construct a complete working environment.

[0009] Preferably, in a possible implementation manner of the first aspect, the generation of the inspection path in step S500 includes: Based on the environmental map, use the A-Star algorithm to plan the starting inspection path, the opposite-side inspection path, and the return flight path; Smooth the starting inspection path, the opposite-side inspection path, and the return flight path through the B-spline algorithm to generate the final inspection path.

[0010] Preferably, in a possible implementation manner of the first aspect, the planning of the starting inspection path is to calculate the point with the shortest distance between the UAV take-off point and the inspection waypoints as the starting point of the starting inspection path, and obtain the starting inspection flight line; The planning of the opposite-side inspection path is to calculate the distances between the end point of the side that has completed the inspection and the head and tail points of the opposite side, and select the point with the shorter distance as the starting point of the opposite-side inspection path to obtain the opposite-side inspection flight line; The planning of the return flight path is to use the completion point of the inspection path on the opposite side as the starting point to obtain the inspection flight path for the return flight.

[0011] Preferably, in a possible implementation manner of the first aspect, the method further includes simultaneously selecting multiple supports for inspection and generating an inspection path for multiple supports according to a path planning algorithm.

[0012] In a second aspect, the present invention provides a bridge support inspection device based on a drone. The device includes the drone itself, the accessory equipment of the drone, a receiving device based on the link of the drone itself, a remote device based on a network connection, a memory, a processor, and a bridge support inspection program based on the drone stored on the memory and executable on the processor, so that the device executes any possible implementation method in the first aspect.

[0013] In a third aspect, the present application provides a computer program product, which may include computer instructions. When the computer instructions run on an electronic device, the electronic device is enabled to execute any possible implementation method in the first aspect.

[0014] The beneficial effects of the present invention are as follows: By integrating technical means such as automatic waypoint generation, path smoothing optimization, high-precision path planning, and complex environment adaptation, an efficient, safe, and accurate bridge inspection solution is provided. The drone can automatically generate and execute precise inspection tasks to ensure precise inspection of structures such as bridge piers, while effectively reducing the risks and costs of manual operations. The present invention not only improves the inspection efficiency but also optimizes the flight stability, ensuring the smooth completion of the inspection task and providing reliable technical support for future intelligent bridge inspections. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0016] Figure 1 This application provides a flowchart of a method for inspecting bridge supports based on a drone.

[0017] Figure 2 This application provides an original point cloud data diagram.

[0018] Figure 3 This application provides a point cloud data diagram after elevation slicing.

[0019] Figure 4This application provides a point cloud data map of each independent pier formed after clustering calculation.

[0020] Figure 5 This application provides a calculated pier contour map.

[0021] Figure 6 This application provides an effect diagram of fitting the pier contour to the original point cloud.

[0022] Figure 7 This application provides an inspection waypoint map.

[0023] Figure 8 This application provides a bridge center line map.

[0024] Figure 9 This application provides an initial inspection path planning map.

[0025] Figure 10 This application provides an inspection path planning map for the opposite side.

[0026] Figure 11 This application provides a return flight path planning map.

[0027] Figure 12 This application provides an overall inspection path map. Detailed implementation manners

[0028] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0029] Embodiment 1: As Figure 1 shown, the present invention provides a method for inspecting bridge bearings based on an unmanned aerial vehicle. The method includes: Step S100: Import the three-dimensional point cloud model of the bridge and the reference slice elevation, perform slicing processing on the point cloud model based on the reference slice elevation, and extract the pier contour.

[0030] In this embodiment, first import the point cloud model of the bridge and the reference slice elevation for extracting the pier contour. Perform slicing processing on the point cloud based on the reference slice elevation to remove invalid points. Perform a minimum rectangular circumscription on the point cloud data after angle and distance screening, so as to obtain the contour of the pier.

[0031] First, obtain the original high-resolution 3D point cloud data of the bridge. In this embodiment, the reference slice elevation of the pier contour is set to 37 - 39m. The camera used for inspection has the following internal orientation elements: the focal length of the camera is 9.1, the sensor width is 7.68, the resolution of the sensor in the horizontal direction is 640, and the resolution in the vertical direction is 512.

[0032] When conducting the inspection work, set the overlap degree in the horizontal direction to 40%, the relative distance to 6m, and the takeoff point coordinates of the UAV to (119.35934, 32.21622, 40.5).

[0033] In terms of path planning parameters, the value of the smooth_factor for route smoothing is 100; in the environmental map parameters, the buffer_distance of the bridge center line is set to 20m, which is used to limit the boundary of path planning; the safe_radius of route planning is set to 1m.

[0034] After that, read and process the point cloud data. Read the LAS file containing bridge information, which contains the 3D point cloud data of the bridge area. The function returns the point cloud data and the metadata of the file (such as coordinate system, point type, etc.). Dilute the read point cloud and save it, and slice it according to the reference elevation for calculating the pier contour input. Figure 2 The original point cloud data is shown as follows. Figure 3 The point cloud data after elevation slicing is shown as follows.

[0035] Next, apply the DBSCAN clustering algorithm to group the point cloud to distinguish the pier from other areas (such as the surrounding ground or buildings). The DBSCAN used in this embodiment is a density-based clustering algorithm that can effectively identify dense areas. By setting the clustering radius (eps) and the minimum number of neighbors (min_samples), the point set belonging to the pier is separated. Figure 4 The point cloud of each independent pier formed after clustering calculation is shown as follows.

[0036] To improve the quality of the point cloud data, remove the noise points, and use the statistical outlier removal (SOR) method for denoising. This method removes the outliers far from most neighbors by calculating the neighborhood statistical information of the points (such as average distance and standard deviation), and finally only retains the valid points in the point cloud data.

[0037] Apply the principal component analysis (PCA) algorithm to the denoised point cloud data to extract the principal component direction of the point cloud data. PCA projects the data onto the most representative direction, thereby screening out the points consistent with the main direction. By analyzing the principal component direction, the basic structural direction of the pier is identified.

[0038] Subsequently, the Open3d library is used to calculate the normal vectors of each point in the point cloud. The normal vectors reflect the orientation of the point cloud surface and can identify whether the point cloud belongs to the surface of the pier. The normal vector of each point is calculated based on its neighboring points and then compared with the principal component directions to filter out the points parallel or perpendicular to the surface of the pier.

[0039] Then, the point cloud after normal vector filtering is further processed. First, by calculating the angle and distance between each point and the centroid, the points with angles close to and the minimum distance from the centroid are retained, and the edge points are removed. By finely filtering the point cloud, only the most representative points are ensured to be retained.

[0040] Finally, the minimum rotated rectangle containing all the filtered points is calculated. The minimum rectangle enclosing can accurately describe the shape of the pier. In this embodiment, the shapely library is used for geometric calculations to obtain the coordinates of the four corner points of the rectangle, and these four corner point coordinates are the pier contour. Figure 5 The calculated pier contour is shown as follows, Figure 6 The effect of fitting the pier contour to the original point cloud is shown as follows.

[0041] Step S200: Calculate the acquisition interval in the horizontal direction according to the load parameters. The load parameters include the relative flight distance, camera focal length, sensor size, and resolution. Determine the center line of the pier based on principal component analysis and use this center line as the inspection baseline.

[0042] In this embodiment, the acquisition interval in the horizontal direction is calculated according to the input interior orientation elements of the load for inspection, and the take-off point of the UAV and the pier support number to be inspected are set. Through the previously calculated pier contour, the center line of the pier is calculated as the object of inspection.

[0043] The system calculates the horizontal interval at each shooting according to the parameters such as the relative distance, camera focal length, sensor size, and resolution given in step S100 to ensure the coverage and overlap of the images during the inspection process. At the same time, this interval distance is used for the generation of subsequent waypoints.

[0044] Step S300: Generate buffer inspection waypoints on both sides of the center line according to the acquisition interval in the horizontal direction, calculate the distances between the UAV take-off point and each waypoint, and select the waypoint with the shortest distance as the starting point of the inspection.

[0045] In this embodiment, the positions of the sampling points are calculated according to the horizontal interval, and buffer is performed on both sides of the center line to form inspection waypoints. Calculate the distances between the UAV starting point and the waypoints, and select the point with the shortest distance as the starting point of the inspection.

[0046] As Figure 7As shown, sampling points are generated according to the center line of the pier parallel to the principal component direction, and inspection waypoints are generated on both sides of the center line according to the input inspection radius using the buffer algorithm. These waypoints are distributed on the left and right sides of the center line, and the specific spacing is calculated based on the horizontal interval distance.

[0047] Step S400: Generate an environmental map including the pier and the preset range around it. The environmental map includes information on the internal structure of the bridge and external obstacles.

[0048] In this embodiment, an environmental map including all piers and the 20-meter range around them is generated, and all areas beyond this range are regarded as obstacles. Combining the information on other internal and external obstacles of the bridge, a complete working environment is constructed.

[0049] As Figure 8 shown, the environmental map is used to implement path planning. The outline of the pier is set as an obstacle, the center points of the paired outlines of the pier are calculated to obtain the center line of the bridge, and the center line of the bridge is set as the boundary that the path planning cannot exceed according to the buffer distance of the bridge center line.

[0050] Step S500: Based on the environmental map, use the A-Star algorithm for path planning and smooth the path through the B-spline algorithm to generate an inspection path.

[0051] In this embodiment, the A-Star algorithm is used for path planning, considering the safety distance (2 meters in this embodiment) to avoid the path being too close to obstacles. Then, the path finding result is smoothed through the B-spline algorithm to ensure that the drone can successfully complete the inspection task. The path starts from the take-off point, passes through the collection waypoints of the pier, and finally returns to the take-off point.

[0052] First, perform the initial inspection path planning. As Figure 9 shown, calculate the point closer to the take-off point of the drone among the inspection points as the starting point of the inspection route, use the A-Star algorithm to plan the path of the drone, and smooth the generated path.

[0053] Next, perform the opposite side inspection path planning. As Figure 10 shown, calculate the distance between the end point of the side that has completed the inspection and the start and end points of the opposite side, and select the point with the shorter distance as the starting point of the opposite side inspection. Similarly, use the A-Star algorithm for path planning and the B-spline algorithm for smoothing to obtain the inspection route of the opposite side.

[0054] Finally, perform the return flight path planning. As Figure 11 shown, use the A-Star algorithm for path planning and the B-spline algorithm for smoothing between the completion point of the opposite side route and the take-off point of the drone to obtain the return flight route.

[0055] The overall inspection path is as Figure 12 shown.

[0056] Embodiment 2: The present invention provides a method for inspecting bridge bearings based on an unmanned aerial vehicle. The method further includes the ability to simultaneously select multiple bearings and perform inspections according to the same path planning algorithm. By adding data on external or other obstacles of the bridge, a more complex operating environment is formed to improve the inspection accuracy and safety.

[0057] Embodiment 3: The present invention provides an inspection device for bridge bearings based on an unmanned aerial vehicle. An inspection device for bridge bearings based on an unmanned aerial vehicle includes the unmanned aerial vehicle itself, the accessory equipment of the unmanned aerial vehicle, a receiving device based on the link of the unmanned aerial vehicle itself, a remote device based on network connection, a memory, a processor, and an inspection program for bridge bearings based on an unmanned aerial vehicle stored on the memory and executable on the processor. When the inspection program for bridge bearings based on an unmanned aerial vehicle is executed by the processor, it implements a method for inspecting bridge bearings based on an unmanned aerial vehicle in Embodiment 1 above.

[0058] Embodiment 4: The present invention further provides a computer program product, including an inspection program for bridge bearings based on an unmanned aerial vehicle. When the inspection program for bridge bearings based on an unmanned aerial vehicle is executed by the processor, it implements a method for inspecting bridge bearings based on an unmanned aerial vehicle as in Embodiment 1 above.

[0059] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these changes and modifications.

Claims

1. A bridge bearing inspection method based on drone, characterized in that: The method comprises: Step S100: importing a three-dimensional point cloud model of a bridge and a reference slice elevation, slicing the point cloud model based on the reference slice elevation, and extracting the pier contour; Step S200: Calculate the horizontal collection interval according to the load parameters, wherein the load parameters include relative flight distance, camera focal length, sensor size and resolution, determine the center line of the pier based on principal component analysis, and use the center line as the inspection baseline; Step S300: generating buffer inspection waypoints on both sides of the center line according to the horizontal collection interval, calculating the distance between the take-off point of the drone and each waypoint, and selecting the closest waypoint as the inspection starting point; Step S400: generating an environment map including the bridge pier and a preset range around it, wherein the environment map includes information on the internal structure of the bridge and external obstacles; Step S500: Based on the environment map, the A-Star algorithm is used for path planning, and the path is smoothed by the B-spline algorithm to generate an inspection path.

2. The bridge bearing inspection method according to claim 1, characterized in that: The extraction of the pier contour in step S100 includes: The three-dimensional point cloud data is subjected to thinning processing, and a bridge pier point set is separated based on a DBSCAN clustering algorithm; The principal direction of the pier is determined by principal component analysis, and the normal vector of each point in the point cloud is calculated to filter out points that are parallel or perpendicular to the pier surface. The minimum rotated rectangle of the filtered point set is extracted, and the coordinates of the four corner points of the rectangle are obtained as the coordinates of the pier contour.

3. The bridge bearing inspection method according to claim 2, characterized in that: The determination of the bridge pier centerline in step S200 includes: Based on the principal component analysis algorithm, the pier point cloud data is projected to the most representative direction to determine the main direction of the pier; The center line of the bridge pier is calculated according to the main direction, and the center line is used as the inspection baseline.

4. The bridge bearing inspection method according to claim 1, characterized in that: The generation of the inspection waypoint in step S300 includes: Generating buffer areas on both sides of the center line according to the horizontal collection interval; Inspection waypoints are generated based on the buffer area, and the distance between each waypoint and the take-off point of the drone is calculated, and the closest waypoint is selected as the inspection starting point.

5. The bridge bearing inspection method according to claim 1, characterized in that: The generation of the environment map in step S400 includes: Based on the three-dimensional point cloud data, an environmental map including the bridge pier and a preset range around it is generated; Areas beyond the preset range are marked as obstacles, and a complete working environment is constructed by combining the internal structure of the bridge and external obstacle information.

6. The bridge bearing inspection method according to claim 1, characterized in that: The generation of the inspection path in step S500 includes: Based on the environmental map, the A-Star algorithm is used to plan the initial inspection path, the opposite inspection path and the return route path; The initial inspection path, the opposite inspection path and the return route path are smoothed by the B-spline algorithm to generate the final inspection path.

7. The bridge bearing inspection method according to claim 6, characterized in that: The planning of the initial inspection path is to calculate the point between the take-off point of the drone and the inspection waypoint that is closest to the inspection waypoint, and use it as the starting point of the initial inspection path to obtain the initial inspection route; The planning of the opposite inspection path is to calculate the distance between the end point of the side that has completed the inspection and the first and last points of the opposite side, select the point with the shorter distance as the starting point of the opposite inspection path, and obtain the inspection route of the opposite side; The return route is planned by taking the completion point of the opposite inspection route as the starting point to obtain the return inspection route.

8. The bridge bearing inspection method according to claim 1, characterized in that: The method also includes selecting multiple supports for inspection at the same time, and generating inspection paths for the multiple supports according to a path planning algorithm.

9. A bridge bearing inspection device based on drone, characterized in that: The device includes the drone itself, the drone's auxiliary equipment, a receiving device based on the drone's own link, a remote device based on a network connection, a memory, a processor, and a drone-based bridge bearing inspection program stored on the memory and executable on the processor. When the drone-based bridge bearing inspection program is executed by the processor, the bridge bearing inspection method described in any one of claims 1 to 8 is implemented.

10. A computer program product, characterized in that The computer program product includes a UAV-based bridge bearing inspection program, and when the UAV-based bridge bearing inspection program is executed by a processor, it implements the bridge bearing inspection method described in any one of claims 1 to 8.

Images