Racing field track local deformation detection method based on LIDAR data

Abstract

The invention discloses a method for detecting local deformation of a racing track of a racing yard based on LIDAR data. The method comprises the following steps: collecting point cloud data of the whole racing yard through a laser radar system; filtering processing is carried out on the collected point cloud data, noise points are removed, ground points and non-ground points are separated, and the data quality is improved; according to data of elevation and reflection intensity, geometric features of the surface of the racing track are extracted, and the local deformation condition of the racing track is preliminarily analyzed; a point is taken in a certain direction of xyz, the point is limited to a certain range to be approximate to a section, local curvature and elevation variation coefficients are calculated, and spatial distribution characteristics of reflection intensity are analyzed to represent deformation conditions of different areas of the racing track; and based on the local curvature, the variable coefficient and the spatial distribution of the reflection intensity, establishing an evaluation system for the local deformation of the racing track. According to the invention, the local deformation condition of the racing track of the racing yard can be efficiently and accurately detected, and a scientific basis is provided for the maintenance of the racing track.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of track detection and maintenance, and particularly relates to a race track local deformation detection method based on LIDAR data. BACKGROUND

[0002] As the core infrastructure of motorsport, the local deformation of the race track is directly related to the driving safety, handling stability, and fairness of the race results. With the rapid development and technological progress of motorsport, the requirements for local deformation of the race track are increasingly high. Local deformation not only affects the grip and handling performance of the car, but also directly relates to the safety of the driver and the smooth progress of the event. Therefore, efficiently and accurately evaluating the local deformation of the race track has become a key technical requirement in track maintenance and event management.

[0003] Traditional methods for detecting local deformation of the race track mainly rely on manual measurement or mechanical equipment-based detection means. Manual measurement methods usually use tools such as rulers and total stations to measure the elevation data of the race track by segment and point, to evaluate the local deformation. However, this method is inefficient and difficult to cover the details of the entire race track, especially on large-scale race tracks, manual measurement not only consumes time and effort, but also introduces human error. Mechanical equipment-based detection methods, such as using local deformation detection vehicles, can improve detection efficiency, but the equipment cost is high, and the coverage range and accuracy of the race track are still limited, making it difficult to meet the high-precision and high-efficiency detection requirements.

[0004] In recent years, with the rapid development of LIDAR technology, its advantages in three-dimensional spatial data acquisition have gradually emerged. LIDAR technology can quickly and accurately acquire large-scale three-dimensional point cloud data, providing a new technical means for race track local deformation detection. However, existing LIDAR data-based race track detection methods are mostly focused on terrain mapping or road detection, lacking customized solutions for the special needs of race track. Race track not only requires high-precision local deformation evaluation, but also needs to consider the geometric characteristics and dynamic changes of the race track. In addition, existing methods have deficiencies in data processing and model optimization, making it difficult to achieve efficient and accurate local deformation evaluation. SUMMARY

[0005] In view of the deficiencies in the prior art, the purpose of the present application is to provide a race track local deformation detection method based on LIDAR data. The traditional race track local deformation detection method relies on manual measurement (such as ruler, level) or mechanical detection vehicle, and has problems such as low efficiency, limited coverage (only local area can be detected), insufficient precision (error > 5mm), and cannot comprehensively evaluate the implicit deformation characteristics of the track surface material wear. The existing detection method based on laser radar (LIDAR) focuses on road or terrain mapping, and lacks professional solutions for high-precision, multi-dimensional deformation analysis of race tracks. The present application breaks through the traditional technical bottleneck and realizes intelligent evaluation of millimeter-level precision, full track coverage and multi-dimensional feature fusion. In order to achieve the above purposes and other advantages according to the present application, a race track local deformation detection method based on LIDAR data is provided, comprising the following steps:

[0006] S1, collecting point cloud data of the entire race track by a laser radar system, including spatial point position, elevation, reflection intensity and other indicators of the track;

[0007] S2, filtering the collected point cloud data to remove noise points, separate ground and non-ground points, and improve data quality;

[0008] S3, extracting the geometric features of the track surface according to the elevation, reflection intensity and other data, and preliminarily analyzing the local deformation of the track;

[0009] S4, taking points in a certain direction of xyz and limiting to a certain range to approximate a section (for example, x±0.1 as a section of x), calculating local curvature (K), elevation coefficient of variation (CVh), and analyzing the spatial distribution characteristics of reflection intensity to represent the deformation of different areas of the track; S5, based on local curvature, coefficient of variation and spatial distribution of reflection intensity, an evaluation system of track local deformation is established to provide data support for track maintenance.

[0010] Preferably, intelligent filtering and feature extraction of LIDAR point cloud data: the present application uses a laser radar system to perform multi-view scanning to obtain high-precision point cloud data of the track, synthesizes complete track point cloud (density > 200 points / ㎡), and the registration error is ≤2mm), obtains spatial point position, elevation, reflection intensity and other information, and removes noise points by adaptive filtering method, eliminates data with too low reflection intensity, and improves data quality. At the same time, through data grouping strategy, taking X-axis direction as characteristic value, the track data is locally segmented, and each section takes a fixed range (x±0.1m) in X-axis, to ensure the continuity and local precision of data analysis.

[0011] Preferably, the multi-dimensional feature fusion local deformation evaluation model is innovative: the elevation, reflectivity, local curvature, coefficient of variation and other multi-dimensional features are innovatively fused into the local deformation evaluation model, and a more comprehensive local deformation evaluation system is constructed.(a) Local curvature calculation: divide the section (range x±0.1m) with X-axis as the reference, adopt KNN algorithm to construct dynamic neighborhood (k=10±3), fit local quadratic surface through moving least squares method (MLS), and solve principal curvature based on eigenvalues of Hessian matrix.(b) Coefficient of variation analysis: calculate the elevation coefficient of variation by using sliding window method (window size=track standard width+20%), set threshold value 0.15 to trigger deformation warning, and quantify surface depression / protrusion degree.(c) Reflectivity spatial distribution: establish local coordinate system through PCA, generate 0.01m resolution grid by using inverse distance weighted (IDW) interpolation, and extract contrast and entropy value indexes by combining gray level co-occurrence matrix (GLCM), when the value>1.5, determine that it is surface material degradation area. This makes the local deformation evaluation change from traditional single elevation analysis to multi-dimensional and intelligent analysis.

[0012] Preferably, compared with the traditional fixed threshold method, the dynamic filtering adapts to the complex reflection environment of the track, and improves the data reliability.

[0013] Preferably, the interpolation accuracy is improved: IDW interpolation is more suitable for non-uniform point cloud distribution than linear interpolation, and the grid resolution reaches 0.01mm, meeting the high-precision requirement of the race track.

[0014] Compared with the prior art, the beneficial effects of the present application are as follows: three-dimensional point cloud data of the track is collected by the LIDAR system, multi-dimensional features such as elevation and reflectivity are combined, geometric features of the track are extracted, and local deformation indexes are calculated. At the same time, through optimization of data processing algorithm and model, efficient and accurate evaluation of local deformation of the track is realized, and reliable data support is provided for track maintenance and event management. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 A flowchart of the LIDAR data-based local deformation detection method of the race track according to the present application;

[0016] Figure 2 An elevation data information processing flowchart of the LIDAR data-based local deformation detection method of the race track according to the present application;

[0017] Figure 3 A reflectivity data information processing flowchart of the LIDAR data-based local deformation detection method of the race track according to the present application;

[0018] Figure 4The fitting result figure of the elevation curve of any section after eliminating the inclination according to the LIDAR data-based local deformation detection method of the racing track of the racing track: the horizontal axis represents the distance of the section along the horizontal direction (unit: meter), and the vertical axis represents the elevation of the corresponding point (unit: meter). The blue scatter points in the figure are the actual elevation distribution of the laser radar point cloud data on the section, and the red curve is the fitting curve obtained based on linear regression. By comparing the deviation of the scatter points and the fitting curve, the local elevation change characteristics and deformation degree of the section can be intuitively reflected. The corresponding section center coordinates of the figure are (331190, 3468523), and the angle is 0°, indicating that the section data as a whole presents a relatively smooth linear trend, and a small amount of scatter points are distributed away from the fitting curve, which is used to represent the local unevenness of the track surface;

[0019] Figure 5 The distance distribution scatter plot of the section elevation to the fitting curve according to the LIDAR data-based local deformation detection method of the racing track of the racing track: the horizontal axis represents the projection coordinates of the track transverse section, and the vertical axis represents the vertical residual distance value of each point to the fitted regression curve. The blue mark in the figure identifies the concave point with a negative residual, and the red mark identifies the convex point with a positive residual. The distribution graph reflects the concave-convex change characteristics of the track cross section. By analyzing the distribution graph, the local convex-concave uneven area of the track surface can be identified, so as to evaluate the track geometry;

[0020] Figure 6 The local elevation standard deviation distribution graph of the track according to the LIDAR data-based local deformation detection method of the racing track of the racing track: the coordinate axes represent the coordinates of X and Y directions in the ground coordinate system. The color in the figure represents the average elevation residual value of the corresponding position, wherein the red color represents the positive deviation area relative to the fitting plane, and the blue color represents the negative deviation area. The spatial distribution graph shows the elevation change law at different positions on the track surface, which can be used to identify the local elevation anomaly or settlement of the track surface, and assist in evaluating the road surface flatness;

[0021] Figure 7 The reflection intensity distribution of any section along the X direction according to the LIDAR data-based local deformation detection method of the racing track of the racing track: the horizontal axis represents the distance along the laser scanning track (unit: meter), and the vertical axis represents the reflection intensity value of the corresponding position. The red scatter points in the figure represent the original reflection intensity data points collected, and the blue curve represents the reflection intensity signal change trend. The green dotted line represents the global average reflection intensity level of all points. The intensity distribution graph can be used to evaluate the uniformity of the track surface material and identify possible deterioration areas, such as areas with significantly reduced reflection intensity, which may correspond to positions of material wear or coating peeling.

[0022] Figure 8A track local three-dimensional curvature distribution map of a race track based on a LIDAR data-based race track local deformation detection method according to the application. DETAILED DESCRIPTION

[0023] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by a person of ordinary skill in the art without creative effort fall within the protection scope of the application.

[0024] REFERENCE Figure 1 A LIDAR data-based race track local deformation detection method, comprising the following steps:

[0025] S1, collecting three-dimensional point cloud data of the entire race track by a laser radar system, and storing the three-dimensional point cloud data as a point cloud data set with a spatial index structure; the point cloud data includes spatial points, reflection intensity, timestamp information, and elevation.

[0026] Further, appropriate LIDAR equipment is selected, and reasonable scanning frequency and accuracy parameters are set.

[0027] Multiple scans are performed to synthesize complete race track point cloud data, so as to reduce data loss caused by measurement blind areas.

[0028] Based on the collected reflection intensity values, possible false measurement points are removed, and the reliability of the data is improved.

[0029] S2, pre-processing the point cloud data, and drawing a track surface deformation three-dimensional heat map to separate ground points and non-ground points; abnormal noise points are removed by an intensity filtering algorithm, the collected point cloud data is filtered and processed to remove noise, and the accuracy and reliability of the data are ensured. The data quality is ensured by filtering and screening according to the reflection intensity data. The track surface point cloud is segmented by a non-ground point separation algorithm and a ground filtering algorithm.

[0030] Further, the method for data processing comprises:

[0031] The mean value and standard deviation of the Intensity column are calculated, the intensity threshold is set, and the points with low reflection intensity are removed to reduce the influence of measurement errors.

[0032] The Z value (elevation) is statistically analyzed, and abnormal data points deviating from the main surface are removed.

[0033] Grouping is performed according to time or spatial position (X coordinate), so that the data at different time or in different regions can be processed respectively, and the calculation efficiency is improved.

[0034] Further, filtering and data processing can be implemented through Python's Pandas, NumPy, and SciPy libraries.

[0035] S3, extract the geometric features of the track surface according to the elevation and reflectivity, and calculate the local deformation index of the track. Specifically, based on the preset grid division method, through data grouping strategy, taking the X-axis direction as the characteristic value, the track data is locally segmented, each section takes a fixed range on the X-axis, such as x0±0.1m, to ensure the continuity and local accuracy of data analysis, and the local curvature and elevation coefficient of variation are calculated in each region as the quantitative indicators of geometric deformation and distribution dispersion.

[0036] Further, the Z value of the screened point cloud data is adjusted, the deviation relative to the average Z value is calculated, the outliers are removed, and the linear regression fitting is used to calculate the distance of each point to the regression line. The deformation degree of the local track surface is evaluated based on the statistical quantity of the distance, such as mean, standard deviation, maximum value and coefficient of variation. The calculation of local curvature includes: based on elevation data, K nearest neighbor (KNN) algorithm is used to construct local neighborhood, quadratic surface is fitted and principal curvature is solved, as a quantitative indicator of geometric deformation. The neighborhood points of each point are found by K nearest neighbor (KNN) algorithm, and the coefficients of the quadratic surface are extracted by fitting the quadratic surface of the point and its neighborhood points. The calculation of the elevation coefficient of variation represents the dispersion degree of the elevation distribution by the ratio of the standard deviation of the region elevation to the average value, and the curvature distribution graph and the fitted surface are displayed through the visualization tool.

[0037] Further, the method for calculating the local deformation index of the track includes:

[0038] Taking the X-axis as the reference, the track is divided into multiple local sections, and the range of each section is set to x±0.1m to ensure the continuity and stability of local analysis;

[0039] K nearest neighbor (KNN) algorithm is used to select 10 nearest neighbor points in each section;

[0040] Least squares method is used to fit the quadratic surface and calculate the principal curvature, which is one of the key indicators for local deformation evaluation;

[0041] The elevation coefficient of variation of the local region is calculated to measure the fluctuation of the track surface height.

[0042] Further, it is realized through Python's scikit-learn library and NumPy library, and the curvature distribution graph is drawn by Matplotlib.

[0043] S4, adopt PCA principal component analysis to extract the main change direction of the track surface, and calculate the index of local curvature and coefficient of variation based on the direction, and further evaluate the local deformation situation combined with the spatial distribution characteristics of the reflection intensity. Specifically: principal component analysis is performed on the point cloud data, and a uniform grid reflection intensity distribution map is generated; the generation of the reflection intensity distribution map includes: performing principal component analysis on the point cloud data, determining the local coordinate system, and generating a uniform grid reflection intensity distribution map through reflection intensity data and interpolation algorithm.

[0044] Further, PCA principal component analysis is adopted, and the coordinates of the point cloud projection to the main direction are calculated. The projected reflection intensity data is sorted by X, and a uniform grid signal (resolution 0.01m) is generated using inverse distance weighted interpolation. The edge value is filled to prevent extrapolation NaN and ensure spatial continuity. Fast Fourier transform is performed on the interpolated reflection intensity signal, the amplitude spectrum is calculated, and the main frequency and its corresponding period are extracted. The periodic fluctuation corresponding to the main frequency can reflect the local deformation of the track surface.

[0045] Further, the more accurate local deformation analysis optimization method includes:

[0046] In each local section, the residual of the Z value (elevation) and the regression surface is calculated, and the standard deviation is calculated to measure the elevation fluctuation of the local area;

[0047] PCA principal component analysis is adopted to extract the main change direction of the track surface, and the coordinates of the point cloud projection to the main direction are calculated;

[0048] Inverse distance weighted (IDW) interpolation is performed on the projected reflection intensity data to improve the calculation stability;

[0049] Combined with the reflection intensity distribution characteristics after interpolation, further assist in local deformation evaluation.

[0050] Further, the spatial distribution analysis of the reflection intensity can be realized through Python Pandas, NumPy and sklearn libraries. PCA is used for main direction extraction, and interpolation method is used to optimize data distribution.

[0051] S5, construct a multi-dimensional evaluation system, fuse local curvature, elevation coefficient of variation and reflection intensity distribution characteristics for analysis, and generate a track deformation grade atlas. The construction method of the multi-dimensional evaluation system includes: constructing a fuzzy comprehensive evaluation model according to the characteristics of local curvature, elevation coefficient of variation and reflection intensity spatial distribution, calculating a comprehensive deformation index, and dividing the deformation grade of the track. The track area is divided into different grades to provide local deformation evaluation results and support track maintenance decision-making.

[0052] The number of devices and processing stages described herein are intended to simplify the description of the application and are not intended to limit the application to the embodiments described. Although embodiments of the application have been disclosed in connection with the specified materials and procedures, it is to be understood that the application is not limited to the disclosed materials and procedures, but rather, the intended application is to suggest that the application is broadly applicable to the pertinent art. Various modifications and alterations of this application will become apparent to those skilled in the art from the disclosure herein, and it is intended that the application shall be limited only by the scope of the claims appended hereto and equivalents thereof.

Claims

1. A method for detecting local deformation of a racetrack based on LiDAR data, characterized in that, The method comprises the following steps: S1, collecting three-dimensional point cloud data of the entire racecourse and storing as a point cloud data set with a spatial index structure; S2, preprocessing the point cloud data and drawing a three-dimensional heat map of the track surface deformation to separate ground points and non-ground points; S3, based on a preset grid division method, regionally segmenting the track and calculating local curvature and elevation coefficient of variation in each region as quantitative indicators of geometric deformation and distribution dispersion; S4, performing principal component analysis on the point cloud data and generating a uniform grid reflectivity distribution map; S5, constructing a multi-dimensional evaluation system, fusing and analyzing local curvature, elevation coefficient of variation, and reflectivity distribution characteristics to generate a track deformation grade map.

2. The method of claim 1, wherein the method further comprises: The point cloud data in step S1 includes spatial point position, reflectivity, timestamp information, and elevation.

3. The method of claim 1, wherein, In step S2, abnormal noise points are removed by an intensity filtering algorithm, and track surface point cloud is segmented by a non-ground point separation algorithm and a ground filtering algorithm.

4. The method of claim 1, wherein, In step S3, the track data is locally segmented by taking the X-axis direction as the characteristic value through a data grouping strategy, and a fixed range is taken on the X-axis, such as x0±0.1m.

5. A method of detecting local deformations of a race track based on LIDAR data according to claim 4, wherein, In step S3, the Z value of the screened point cloud data is adjusted, the deviation of each point relative to the average Z value is calculated, and after removing the abnormal values, linear regression fitting is used to calculate the distance of each point to the regression line, and the deformation degree of the local pavement is evaluated based on the statistical quantity of the distance.

6. A method of detecting local deformations of a race track based on LIDAR data according to claim 5, wherein, In step S3, the calculation of local curvature includes: based on the elevation data, a local neighborhood is constructed by K nearest neighbor (KNN) algorithm, a quadratic surface is fitted and principal curvature is solved as a quantitative indicator of geometric deformation; the calculation of the elevation coefficient of variation is represented by the ratio of the standard deviation of the elevation in the point cloud data in the corresponding local region to the average value, which represents the dispersion degree of the elevation distribution.

7. The method of claim 1, wherein, In step S4, the generation of the reflectivity distribution map includes: performing principal component analysis on the point cloud data, determining the local coordinate system, and generating a uniform grid reflectivity distribution map through reflectivity data and interpolation algorithm.

8. The method of claim 1, wherein, In step S5, the construction method of the multi-dimensional evaluation system includes: according to the characteristics of local curvature, elevation coefficient of variation, and reflectivity spatial distribution, a fuzzy comprehensive evaluation model is constructed, a comprehensive deformation index is calculated, and the deformation grade of the track is divided.

Images