Method and system for quickly repairing pavement cracks

Abstract

The invention discloses a pavement crack rapid repairing method and system, and the method comprises the steps: obtaining multi-source data of a target pavement through a multi-source device, and carrying out the preprocessing of the multi-source data; wherein the multi-source data comprises a target road surface image and target road surface point cloud data; outputting crack features of the target pavement according to the preprocessed target pavement image and the target pavement point cloud data; wherein the crack characteristics comprise the crack length, the crack width, the crack depth, the crack trend, the crack branch number and the crack position; generating a pavement crack repairing scheme based on the crack characteristics; wherein the repairing scheme comprises a repairing material, a repairing mode and repairing process parameters; according to the crack position and the repairing scheme, a repairing track of repairing equipment is determined; and according to the repairing track and the repairing scheme, the repairing equipment is controlled to move to the target position to quickly and accurately repair the pavement crack, so that the repairing efficiency is improved, the cost is reduced, and the repairing quality is ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of road surface repair, in particular to a road surface crack rapid repair method and system. BACKGROUND

[0002] With the continuous increase of traffic volume, due to road aging, wheel pressure, temperature change, natural disasters and other factors, cracks and potholes gradually appear on the road surface, if not repaired in time, not only will affect the service life of the road and driving safety, but also will lead to crack expansion, road structure damage, and even cause traffic safety hazards.

[0003] At present, the traditional road crack repair method is mostly manual repair, which has the problems of low repair efficiency, unstable quality and difficult precise control in the repair process. SUMMARY

[0004] In order to overcome the shortcomings of the prior art, the present application provides a road crack rapid repair method and system to solve the problems in the prior art.

[0005] To achieve the above purpose, the present application provides the following technical scheme: a road crack rapid repair method, comprising: Obtain multi-source data of the target road surface through a multi-source device, and preprocess the multi-source data; wherein the multi-source data includes target road surface image and target road surface point cloud data; According to the preprocessed target road surface image and target road surface point cloud data, output the crack characteristics of the road surface crack of the target road surface; wherein the crack characteristics include crack length, crack width, crack depth, crack direction, crack branch number and crack position; Generate a repair scheme of the road surface crack based on the crack characteristics; wherein the repair scheme includes repair material, repair method and repair process parameters; According to the crack position and the repair scheme, determine the repair trajectory of the repair device; Control the repair device to move along the repair trajectory, and repair the road surface crack based on the repair scheme.

[0006] In one of the embodiments, according to the preprocessed target road surface image and target road surface point cloud data, output the crack characteristics of the road surface crack of the target road surface, comprising: Crack segmentation is performed on the preprocessed target road surface image to generate a binary mask M(x, y) of the crack region; wherein M(x, y) = 1 represents the crack region, and M(x, y) = 0 represents the non-crack region; According to the crack binary mask and the target road surface point cloud data, determine the crack characteristics.

[0007] In one of the embodiments, the crack feature is determined according to the crack binary mask and the target road surface point cloud data, including: Based on the crack binary mask, the crack center line is extracted by a skeletonization algorithm, the length of the center line pixels is calculated, and the pixel length is converted into actual length to obtain the crack length; The cross section in the vertical direction along the crack center line is taken, the distance of the edges on both sides of the crack is calculated, and the crack width is determined based on the average width; Based on the target road surface point cloud data, the point cloud height difference in the crack area is extracted to calculate the crack depth; The direction vector of the crack center line is determined by principal component analysis to determine the crack direction; Based on the crack binary mask, the number of intersection points of the skeletonization is detected to determine the number of crack branches; The position of the crack center point in the image coordinate system is determined, and the crack position is mapped to the world coordinate system through the target road surface point cloud data.

[0008] In one of the embodiments, the repair scheme of the road surface crack is generated based on the crack feature, including: The repair material is determined according to the crack width and the crack depth; The repair method is determined according to the crack depth and the number of crack branches; The repair process parameters are determined according to the crack volume, nozzle flow rate and material prediction; The repair scheme is generated according to the repair material, repair method and repair process parameters.

[0009] In one of the embodiments, the repair material is determined according to the crack width and the crack depth, including: If the crack width < 3mm and the crack depth < 5mm, the repair material is determined to be low-viscosity epoxy resin; If 3mm ≤ the crack width < 10mm, or 5mm ≤ the crack depth < 20mm, the repair material is determined to be polyurethane sealant; If the crack width ≥ 10mm, or the crack depth ≥ 20mm, the repair material is determined to be asphalt-based filling material.

[0010] In one of the embodiments, the repair method is determined according to the crack depth and the number of crack branches, including: If the number of crack branches = 0, the repair method is determined to be continuous linear filling, and the nozzle moves along the crack center line; If the number of crack branches > 0, the repair method is determined to be segmented filling, the main crack is processed first, and then the branches are filled. if the number of crack branches > preset number, determining that the repair method is to use grid repair to cover the branch area; if the crack depth D > 50mm, determining that the repair method is to first perform bottom filling and then surface sealing.

[0011] In one embodiment, the repair process parameters are determined according to the crack volume, nozzle flow rate and material prediction, including: the crack volume is determined according to the crack length, crack width and crack depth; the nozzle flow rate is determined according to the crack width and the material viscosity of the repair material; the material prediction is determined according to the crack volume and the material density of the repair material.

[0012] In one embodiment, the crack center line is discretized into a series of path points, and a path algorithm is used to generate a smooth path from the starting point to the ending point; the path points are smoothed using a spline curve, and the path points in the world coordinate system are converted into the local coordinate system of the nozzle of the repair device to obtain a nozzle motion trajectory; based on the nozzle motion trajectory, a repair trajectory of the repair device is generated.

[0013] In one embodiment, the repair device is controlled to move along the repair trajectory, and the pavement crack is repaired based on the repair scheme, including: the initial position and attitude of the repair device are initialized, and the alignment of the nozzle and the crack position is calibrated; the repair device is driven to move along the repair trajectory using a PID controller, so that the nozzle of the repair device moves along the nozzle motion trajectory; the nozzle is controlled to release the repair material according to the repair scheme.

[0014] A pavement crack rapid repair system, comprising: an acquisition module for acquiring multi-source data of a target pavement through a multi-source device and pre-processing the multi-source data; wherein the multi-source data includes target pavement images and target pavement point cloud data; an output module for outputting crack features of pavement cracks of the target pavement according to the pre-processed target pavement images and target pavement point cloud data; wherein the crack features include crack length, crack width, crack depth, crack direction, number of crack branches and crack position; a generation module for generating a repair scheme of the pavement cracks based on the crack features; wherein the repair scheme includes repair material, repair method and repair process parameters; A determining module is configured to determine a repair trajectory of a repair device according to the crack position and the repair scheme. A control module is configured to control the repair device to move along the repair trajectory and repair the pavement crack based on the repair scheme.

[0015] Compared with the prior art, the present application has the beneficial effects that: The present application integrates crack identification, crack classification, path planning and material spraying into a closed-loop system through the establishment of a closed-loop repair strategy of "crack identification under multi-source data-crack classification-dynamic path planning-precise material spraying", thereby realizing efficient and precise pavement repair to improve repair efficiency, reduce cost and ensure repair quality. Through multi-source data, the morphology and characteristics of pavement cracks are quickly and accurately identified to improve the accuracy and efficiency of crack identification, reduce manual intervention, and according to crack classification, the material distribution and spraying strategy are optimized to realize intelligent selection of repair materials and processes, reduce material waste, and improve repair quality and durability. At the same time, through dynamic path planning, the complex crack morphology is adapted to improve repair efficiency, shorten construction time, meet the demand for rapid traffic, reduce repair cost, and be suitable for various crack types and pavement scenes. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 A flowchart of a pavement crack rapid repair method provided by the present application is shown. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0018] As shown in Figure 1 , a pavement crack rapid repair method provided by the present application comprises: S1, acquiring multi-source data of a target pavement by a multi-source device and pre-processing the multi-source data; wherein the multi-source data comprises target pavement images and target pavement point cloud data; comprising: S11, moving an image acquisition device (such as an industrial camera installed on a drone or a mobile robot) along the pavement at a preset interval time to generate a continuous image sequence , t is a time stamp. Ensure that the image resolution is at least 1920x1080, and the format is RGB or grayscale S12. Synchronously record point cloud data using 3D point cloud acquisition equipment (such as a laser scanner or structural scanner mounted on a drone or mobile robot). For spatial coordinates, For timestamps. Point cloud resolution is controlled within 0.5cm.

[0019] S13. Align the target road surface image and the target road surface point cloud data using timestamps. Data alignment ensures temporal and spatial consistency between the two. GPS / INS systems can be used to assist in positioning.

[0020] S14. Preprocess the target road surface image and target road surface point cloud data to obtain the preprocessed target road surface image. and target road surface point cloud data For spatial coordinates, For timestamps, images The format is either RGB or grayscale. Preprocessing includes: The target road surface image is processed with denoising, distortion correction, and / or polarization filtering to improve crack segmentation accuracy, eliminate illumination effects, reduce environmental noise interference with crack detection, avoid road surface reflections interfering with crack identification, and prevent affecting subsequent image-point cloud registration. For example, Gaussian filtering (kernel size 5×5, standard deviation σ=1.5) is applied to remove noise, as shown in the following formula: ;in, The filtered image pixel values ​​are located at the image coordinates. , These are the original image pixel values. For the original image in coordinates pixel values, The standard deviation of the Gaussian kernel controls the smoothness of the filter. These are the relative coordinates of the Gaussian kernel, the range of which is determined by the kernel size (e.g., ...). nuclear, ), These are the normalization coefficients of the Gaussian function, ensuring that the sum of the weights is 1. The exponential function is used; the Brown-Conrady distortion model is used to perform distortion correction on the target road surface image to eliminate image distortion, improve crack segmentation accuracy, and avoid affecting subsequent image-point cloud registration; Stokes vector polarization imaging is used to perform bias vibration filtering on the target road surface image to eliminate illumination effects, improve crack segmentation accuracy, and avoid road surface reflection interfering with crack identification; adaptive histogram equalization (CLAHE) can also be used to enhance the target road surface image to improve local image contrast.

[0021] The target road surface point cloud data is ground segmented and / or outlier removed, so as to extract the road surface point cloud and remove the noise. For example, the ground segmentation is performed by PANSAC plane fitting, and the outlier removal is performed by statistical filtering. The voxel size filter (voxel size = 1 cm) is used to reduce the point cloud density and reduce the calculation amount; the statistical filter (k = 50, threshold = 1.0) is used to remove outliers.

[0022] The target road surface image pixel and the point cloud coordinate mapped by the target road surface point cloud data are obtained to obtain the preprocessed target road surface image and the target road surface point cloud data. Specifically, the target road surface image pixel is mapped to the target road surface point cloud coordinate by a camera-LiDAR extrinsic matrix T (4x4 transformation matrix): , wherein, is a three-dimensional coordinate in a world coordinate system, is a target road surface image pixel coordinate, is a 4x4 extrinsic transformation matrix containing rotation and translation components, obtained by camera-LiDAR calibration, is a homogeneous representation of a target road surface image coordinate, is a homogeneous representation of a world coordinate.

[0023] According to the need, the image acquisition device and the 3D point cloud acquisition device can be arranged on the unmanned aerial vehicle, and the unmanned aerial vehicle is controlled to realize efficient road surface detection. The positioning module and the working state real-time monitoring module are further arranged on the unmanned aerial vehicle. The position information and the real-time working state of the unmanned aerial vehicle are obtained in real time, so that the phenomenon of “fishing” and “stealing” of the staff during work can be effectively avoided, and the information can be used as reference information for subsequent processing of multi-source data.

[0024] In the embodiment, the target road surface is efficiently inspected and the road surface crack is accurately detected by combining the multi-source data, so that the problems of low efficiency, high omission rate and hysteresis of the traditional repair method which depends on manual detection are solved. The phenomenon of low efficiency or “stealing” of the manual inspection is avoided, and the crack is not found until the middle or later stage. It should be noted that the target road surface can be surrounded before the target road surface is detected to warn the passing vehicles.

[0025] S2, according to the preprocessed target road surface image and the target road surface point cloud data, the crack characteristics of the target road surface crack are output; wherein, the crack characteristics include crack length, crack width, crack depth, crack direction, crack branch number and crack position; specifically, including: S21, the target road surface image is crack segmented to generate a binary mask M(x, y) of the crack region; wherein, M(x, y) = 1 represents the crack region, and M(x, y) = 0 represents the non-crack region; including: Convert the RGB image to a grayscale image I_gray(x, y) = 0.299R + 0.587G + 0.114B; Perform crack segmentation using a pre-trained U-Net model (based on ResNet-50 backbone network, training dataset is a public crack dataset such as Crack500) to generate a binary mask: ; Apply Canny edge detection (low threshold 50, high threshold 150, Sobel kernel 3x3) to extract crack boundaries and further refine the mask ; Use morphological operations (such as opening operation, kernel size 3x3) to remove false crack areas.

[0026] S24, determine the crack features based on the crack binary mask and the target road point cloud data. Specifically, it includes: S241, based on the crack binary mask, extract the crack centerline by skeletonization algorithm, calculate the centerline pixel length, and convert the pixel length to actual length to get the crack length; specifically: Apply Zhang's thinning algorithm to convert the crack binary mask to a single-pixel-wide skeleton ; Calculate the skeleton pixel length : ; wherein, is the pixel length of the crack skeleton in the image, is the pixel set of the crack skeleton, generated by Zhang's thinning algorithm, is the pixel point coordinate on the skeleton.

[0027] Convert to actual length by camera intrinsic parameters (focal length f, sensor size s): ; wherein, is the actual length, is the distance from the camera to the road, is the camera focal length (pixels), is the sensor pixel size (m / pixel).

[0028] S242, take the vertical section along the crack centerline, calculate the distance of the edges on both sides of the crack, and determine the crack width based on the average width; specifically: Uniformly sample M points along the skeleton Calculate the edge distance on the vertical section of each point: ; wherein, is the width of the crack on the vertical section of the jth sampling point; , is the coordinate of the left edge point on the vertical section of the jth point of the crack skeleton (pixels), is the distance from the camera to the road, is the focal length of the camera (pixels), is the pixel size of the sensor (m / pixel).

[0029] Average width is calculated as: and the average width is determined as the crack width. Wherein, is the average width of the crack, is the total number of vertical section sampling points (usually 50-100).

[0030] S243, based on the target road point cloud data, extracting point cloud height difference in the crack area, calculating the crack depth; specific: extracting point cloud subset in the crack area ; ; calculating the depth: ; ; wherein, is the depth of the crack, is the z coordinate of the point cloud point in the crack area; two-dimensional coordinates of the point cloud point, which needs to meet the crack binary mask condition ; is the binary mask of the crack area (1 represents crack, 0 represents non-crack).

[0031] S244, determine the direction vector of the crack center line by principal component analysis, determine the crack direction; specific: applying PCA to the point cloud subset extract the principal direction vector: V = eigenvector(cov( )); wherein, is the crack direction vector (two-dimensional unit vector), cov( ) is the covariance matrix of the point cloud subset of the crack area, eigenvector( ) is the maximum eigenvalue of the covariance matrix, is the point cloud subset of the crack area, containing only =1 points, is the original point cloud data, containing the coordinates of all points .

[0032] S245, based on the crack binary mask, detecting the number of intersection points of the skeleton, determining the number of crack branches; specific: detecting the intersection points (pixel connection number ≥3) in the skeleton S(x, y), counting the number of branches .

[0033] S246, determine the position of the crack center point in the image coordinate system, and map it to the world coordinate system through the 3D point cloud to determine the crack position. Specifically: Calculate the centroid of the skeleton S(x, y): ; wherein, is the centroid of the crack skeleton in the image coordinate system (unit: pixel), is the pixel coordinate in the crack skeleton S(x, y), is the total number of pixels of the crack skeleton; Map to the world coordinate through the extrinsic matrix T: . Wherein, is the position of the crack centroid in the world coordinate system (unit: meter), is the extrinsic transformation matrix.

[0034] According to the need, after determining the crack feature, the consistency of the image segmentation result and the point cloud height change can be compared, and if an abnormality (such as a false crack) is found, the segmentation threshold is adjusted again; and a random forest or SVM model is used to classify the crack features, and non-crack areas (such as road surface texture) are removed, and the crack features are output: .

[0035] S3, generate a repair scheme for the pavement crack based on the crack feature; wherein the repair scheme includes repair material, repair method and repair process parameters; specifically, including: S31, determine the repair material according to the crack width and crack depth; specifically, including: S311, if the crack width < 3mm and the crack depth < 5mm, determine the repair material as low viscosity epoxy resin (viscosity < 500mPa·s); S312, if 3mm ≤ crack width < 10mm, or 5mm ≤ crack depth < 20mm, determine the repair material as polyurethane sealant (tensile strength > 2MPa); S313, if the crack width ≥ 10mm, or the crack depth ≥ 20mm, determine the repair material as asphalt-based filling material (compressive strength > 10MPa).

[0036] S32, determine the repair method according to the crack depth and the number of crack branches; specifically, including: S321, if the number of crack branches = 0, determine the repair method as continuous linear filling, and move the nozzle along the crack center line; S322, if the number of crack branches > 0, determine the repair method as segmented filling, first process the main crack, and then fill the branches; S323. If the number of crack branches is greater than the preset number, the repair method is determined to be a mesh repair to cover the branch area; S324. If the crack depth D>50mm, the repair method is to first fill the bottom layer and then seal the surface.

[0037] It should be noted that if the crack depth D > 50mm, a combination of two repair methods should be used. For example, when the crack depth D > 50mm and the number of crack branches is 0, the repair method is to first fill the bottom layer with a continuous linear filler, and then seal the surface. This ensures the stability and quality of the repair.

[0038] S33. Determine the repair process parameters based on the crack volume, nozzle flow rate, and predicted material quantity; specifically, including: S331. Determine the crack volume based on the crack length, crack width, and crack depth; specifically: ,in, For the crack volume, The crack width (i.e., the average width of the crack). This represents the crack depth.

[0039] S332. Determine the nozzle flow rate based on the crack width and the viscosity of the repair material; specifically: set the nozzle flow rate based on the crack width and the material viscosity. (ml / s), usually = 10-50ml / s. The material flowability is ensured by setting the nozzle pressure P (0.1-0.5MPa) and the material heating temperature T (20-60°C).

[0040] S333. Determine the predicted material quantity based on the crack volume and the material density of the repair material. Specifically: ;in, For the predicted quantity of materials, For material density (such as epoxy resin) ), Loss factor (usually taken as 0.1-0.2, representing 10%-20% loss).

[0041] If needed, particle swarm optimization (PSO) algorithms can also be used to optimize material usage and repair time. ;in, For the predicted quantity of materials, , Weighted by material cost and time cost, Output the total repair time; output the repair plan.

[0042] S34. Generate a repair plan based on the repair materials, repair methods, and repair process parameters.

[0043] In the present embodiment, since the amount of repair material in the existing repair method depends on empirical estimation, the error rate is >30%, and the amount of repair material is estimated by experience, which is easy to cause waste or uneven repair and insufficient precision; therefore, the present application determines the repair scheme according to the repair material, repair method and repair process parameters, which can effectively solve the above problems, and at the same time avoid the use of high-cost materials or the same material for filling different cracks, which leads to excessive repair problems and resource waste.

[0044] S4, according to the crack position and the repair scheme, determining the repair trajectory of the repair equipment; specifically, including: S41, discretizing the crack center line into a series of path points, and using path algorithm to generate a smooth path from the starting point to the ending point; Specifically: discretizing the crack skeleton S(x,y) into equidistant path points ; interval cm, using point cloud data to determine the z coordinate of the path point, ensuring that the trajectory fits the road surface height. Using A* algorithm to generate the shortest path, considering obstacles (such as road bumps): f(n)=g(n)+h(n) where f(n) is the total cost of node n. g(n) is the path length from the starting point to n. h(n) is the estimated distance (Euclidean distance) from n to the end point. Ensure that the path meets the device constraints (such as minimum turning radius R_min =10cm).

[0045] S42, using spline curve to smooth the path points, specifically: applying cubic B-spline curve to smooth the path: , where, is the interpolated path point coordinate (x,y,z), k-order B-spline basis function, represents the control point coordinate , control curvature rate, keep the device motion smooth, represents the parameterized variable, n represents the total number of control points, represents the order of B-spline; and calculate the total length of the path, specifically: . Where, is the jth path point coordinate, is the total number of path points.

[0046] S43, and convert the path points in the world coordinate system into the local coordinate system of the nozzle of the repair equipment to obtain the nozzle motion trajectory. Specifically, convert the world coordinate system path into the local coordinate system of the nozzle: . Where, a path point (x, y, z) in a local coordinate system of the nozzle, a 4x4 transformation matrix of the device pose, including rotation and translation, a path point (x, y, z) in a world coordinate system.

[0047] S44, generating a repair trajectory of the repair device based on the nozzle movement trajectory.

[0048] It needs to be understood that in order to ensure that the nozzle can stably move along the nozzle movement trajectory, the repair trajectory of the repair device needs to be generated according to the nozzle movement trajectory at this time, so as to ensure the stability and accuracy of the repair work. Specifically, the nozzle movement trajectory is corrected according to the position of the nozzle and the position of the moving wheel of the repair device, and then the repair trajectory of the repair device is obtained.

[0049] S5, controlling the repair device to move along the repair trajectory and repairing the pavement crack based on the repair scheme. Specifically, it includes: S51, initializing the initial position and attitude of the repair device and aligning the nozzle and the crack position; S52, using a PID controller to drive the repair device to move along the repair trajectory, so that the nozzle of the repair device moves along the nozzle movement trajectory; S53, controlling the nozzle to release the repair material according to the repair scheme.

[0050] In this embodiment, in the traditional repair method, the repair mechanical path is fixed and lacks dynamic path planning, which is difficult to adapt to the changes of different crack shapes and pavement environments, resulting in low automation and poor repair precision. Therefore, according to the crack position and the repair scheme, the repair trajectory of the repair device is determined, and the repair device is controlled to repair the pavement crack quickly according to the repair trajectory and the repair scheme, which can effectively solve the above problems.

[0051] The application establishes a closed-loop repair strategy of "crack identification under multi-source data-crack grading-dynamic path planning-precise material spraying", integrates crack identification under multi-source data, crack grading, path planning and material spraying into a closed-loop system, and realizes efficient and accurate pavement repair to improve repair efficiency, reduce cost and ensure repair quality. Through multi-source data, the shape and characteristics of the pavement crack are quickly and accurately identified, the accuracy and efficiency of crack identification are improved, manual intervention is reduced, and according to the crack grading, the material distribution and spraying strategy are optimized, the intelligent selection of repair materials and process is realized, the material waste is reduced, and the repair quality and durability are improved. At the same time, through dynamic path planning, the complex crack shape is adapted, the repair efficiency is improved, the construction time is shortened, the rapid traffic demand is met, the repair cost is reduced, and it is suitable for various crack types and pavement scenes.

[0052] A road surface crack rapid repair system comprises: An acquisition module is configured to acquire multi-source data of a target road surface through a multi-source device and to pre-process the multi-source data, wherein the multi-source data comprises a target road surface image and target road surface point cloud data; An output module is configured to output crack features of a road surface crack of the target road surface according to the pre-processed target road surface image and target road surface point cloud data, wherein the crack features comprise crack length, crack width, crack depth, crack direction, crack branch quantity and crack position; A generation module is configured to generate a repair scheme of the road surface crack based on the crack features, wherein the repair scheme comprises a repair material, a repair method and repair process parameters; A determination module is configured to determine a repair trajectory of a repair device according to the crack position and the repair scheme; A control module is configured to control the repair device to move along the repair trajectory and to repair the road surface crack based on the repair scheme.

[0053] By collecting road surface crack images and three-dimensional topography, crack width, depth, direction, branch quantity and other features are extracted, the crack is divided into four levels (micro crack / medium crack / wide crack / net crack) according to the features, the material type and the amount coefficient are output, the material amount of the crack is dynamically calculated, the spraying angle is optimized in real time based on the motion trajectory and obstacle avoidance, and the repair flatness is detected by an infrared sensor, data is fed back to the hierarchical model for iterative optimization, thereby effectively solving the problem that traditional road surface crack repair mainly relies on manual operation or semi-automatic equipment, improving the efficiency and precision of road surface crack repair, combining crack identification, crack classification, dynamic path planning, material precise spraying and other technical means under multi-source data, solving the problems of low efficiency, poor precision and unstable quality in the existing road surface repair method, and realizing rapid and accurate repair of road surface cracks.

[0054] The foregoing descriptions of specific exemplary embodiments of the application are for the purpose of illustration and example only and are not intended to be limiting of the application which is defined solely by the claims appended hereto and equivalents thereof. Numerous modifications, changes, variations, substitutions, and equivalents will occur to those skilled in the art upon reading this specification and the teachings of the application provided herein. It is intended that the application be construed as including all such modifications, changes, variations, substitutions, and equivalents.

Claims

1. A method for rapid repair of road surface cracks, characterized in that, include: Multi-source data of the target road surface is acquired through multi-source devices, and the multi-source data is preprocessed; wherein, the multi-source data includes target road surface images and target road surface point cloud data; Based on the preprocessed target road surface image and target road surface point cloud data, the crack features of the road surface cracks of the target road surface are output; wherein, the crack features include crack length, crack width, crack depth, crack direction, number of crack branches and crack location. A repair plan for the road surface crack is generated based on the crack characteristics; wherein, the repair plan includes repair materials, repair methods, and repair process parameters; Based on the location of the crack and the repair plan, the repair trajectory of the repair equipment is determined; The repair equipment is controlled to move along the repair trajectory, and the road surface cracks are repaired based on the repair plan.

2. The method for rapid repair of road surface cracks according to claim 1, characterized in that, Based on the preprocessed target road surface image and target road surface point cloud data, the crack features of the road surface cracks in the target road surface are output, including: Crack segmentation is performed on the preprocessed target road surface image to generate a binary mask M(x, y) for the crack region; where M(x,y)=1 represents the crack region and M(x,y)=0 represents the non-crack region. The crack features are determined based on the crack binarization mask and the target road surface point cloud data.

3. The method for rapid repair of road surface cracks according to claim 2, characterized in that, Based on the crack binarization mask and the target road surface point cloud data, the crack features are determined, including: Based on the aforementioned crack binarization mask, the crack centerline is extracted using a skeletonization algorithm, the pixel length of the centerline is calculated, and the pixel length is converted into the actual length to obtain the crack length. Take a section perpendicular to the centerline of the crack, calculate the distance between the two edges of the crack, and determine the crack width based on the average width. Based on the point cloud data of the target road surface, the point cloud height difference is extracted within the crack area, and the crack depth is calculated. Principal component analysis is used to determine the direction vector of the crack centerline and thus the crack orientation. Based on the crack binarization mask, the number of skeletonized intersections is detected to determine the number of crack branches; The location of the crack center point in the image coordinate system is determined, and the crack location is determined by mapping the target road surface point cloud data to the world coordinate system.

4. The method for rapid repair of road surface cracks according to claim 1, characterized in that, Based on the crack characteristics, a repair plan for the pavement cracks is generated, including: The repair material is determined based on the crack width and the crack depth; The repair method is determined based on the crack depth and the number of crack branches; The repair process parameters are determined based on the crack volume, nozzle flow rate, and predicted material quantity. The repair plan is generated based on the repair materials, repair methods, and repair process parameters.

5. A method for rapid repair of road surface cracks according to claim 4, characterized in that, Based on the crack width and crack depth, the repair material is determined, including: If the crack width is less than 3 mm and the crack depth is less than 5 mm, the repair material is determined to be low-viscosity epoxy resin. If 3mm ≤ crack width < 10mm, or 5mm ≤ crack depth < 20mm, the repair material is determined to be polyurethane sealant; If the crack width is ≥10mm or the crack depth is ≥20mm, the repair material is determined to be an asphalt-based filler material.

6. A method for rapid repair of road surface cracks according to claim 4, characterized in that, Based on the crack depth and the number of crack branches, the repair method is determined, including: If the number of crack branches is 0, the repair method is determined to be continuous linear filling, and the nozzle moves along the center line of the crack. If the number of crack branches is greater than 0, the repair method is determined to be segmented filling, which involves first treating the main crack and then filling the branches. If the number of crack branches is greater than the preset number, the repair method is determined to be a grid-like repair to cover the branch area; If the crack depth D > 50 mm, the repair method is determined to be to first fill the bottom layer and then seal the surface.

7. A method for rapid repair of road surface cracks according to claim 4, characterized in that, Based on the crack volume, nozzle flow rate, and predicted material quantity, the repair process parameters are determined, including: The crack volume is determined based on the crack length, crack width, and crack depth; The nozzle flow rate is determined based on the crack width and the material viscosity of the repair material; The predicted amount of material is determined based on the crack volume and the material density of the repair material.

8. A method for rapid repair of road surface cracks according to claim 1, characterized in that, Based on the location of the crack and the repair plan, the repair trajectory of the repair equipment is determined, including: The crack centerline is discretized into a series of path points, and a path algorithm is used to generate a smooth path from the start point to the end point. The path points are smoothed using spline curves, and the path points in the world coordinate system are converted into the local coordinate system of the nozzle of the repair device to obtain the nozzle motion trajectory. Based on the nozzle movement trajectory, the repair trajectory of the repair device is generated.

9. A method for rapid repair of road surface cracks according to claim 1, characterized in that, Controlling the repair equipment to move along the repair trajectory and repairing the road surface cracks based on the repair plan includes: Initialize the initial position and orientation of the repair device, and calibrate the alignment of the nozzle with the crack location; The repair equipment is driven to move along the repair trajectory using a PID controller, so that the nozzle of the repair equipment moves along the nozzle movement trajectory. According to the repair plan, the nozzle is controlled to release the repair material.

10. A rapid road surface crack repair system, characterized in that, include: The acquisition module is used to acquire multi-source data of the target road surface through multi-source devices and preprocess the multi-source data; wherein, the multi-source data includes target road surface images and target road surface point cloud data; The output module is used to output the crack features of the road surface cracks of the target road surface based on the preprocessed target road surface image and target road surface point cloud data; wherein, the crack features include crack length, crack width, crack depth, crack direction, number of crack branches and crack location. A generation module is used to generate a repair plan for the pavement crack based on the crack characteristics; wherein the repair plan includes repair materials, repair methods, and repair process parameters; The determination module is used to determine the repair trajectory of the repair equipment based on the location of the crack and the repair plan; The control module is used to control the repair equipment to move along the repair trajectory and to repair the road surface cracks based on the repair plan.

Images