CN115222916A - Non-contact type road surface friction coefficient detection method - Google Patents
Non-contact type road surface friction coefficient detection method Download PDFInfo
- Publication number
- CN115222916A CN115222916A CN202210847074.3A CN202210847074A CN115222916A CN 115222916 A CN115222916 A CN 115222916A CN 202210847074 A CN202210847074 A CN 202210847074A CN 115222916 A CN115222916 A CN 115222916A
- Authority
- CN
- China
- Prior art keywords
- road surface
- friction coefficient
- index set
- detection method
- dimensional
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 26
- 238000013507 mapping Methods 0.000 claims abstract description 20
- 238000012360 testing method Methods 0.000 claims abstract description 15
- 238000010276 construction Methods 0.000 claims abstract description 6
- 238000004422 calculation algorithm Methods 0.000 claims description 16
- 239000013598 vector Substances 0.000 claims description 16
- 230000006870 function Effects 0.000 claims description 7
- 238000000611 regression analysis Methods 0.000 claims description 6
- 238000007621 cluster analysis Methods 0.000 claims description 5
- 230000005540 biological transmission Effects 0.000 claims description 3
- 238000000034 method Methods 0.000 abstract description 20
- 238000005259 measurement Methods 0.000 abstract description 5
- 238000006073 displacement reaction Methods 0.000 description 10
- 238000010586 diagram Methods 0.000 description 7
- 238000001914 filtration Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 6
- 238000005070 sampling Methods 0.000 description 5
- 238000011161 development Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 230000000875 corresponding effect Effects 0.000 description 3
- 230000001419 dependent effect Effects 0.000 description 3
- 230000033001 locomotion Effects 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 239000002245 particle Substances 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 238000007476 Maximum Likelihood Methods 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- -1 accumulated snow Substances 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 239000010426 asphalt Substances 0.000 description 1
- 239000004568 cement Substances 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000001035 drying Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012876 topography Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
- 238000009736 wetting Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/20—Finite element generation, e.g. wire-frame surface description, tesselation
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
-
- G06T5/70—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
- G06T7/344—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/40—Analysis of texture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/762—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using clustering, e.g. of similar faces in social networks
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Abstract
The invention discloses a non-contact road surface friction coefficient detection method, which comprises the following steps: three-dimensional texture scanning is carried out on the road surface through laser construction depth instrument measurement and a visible light camera to obtain a plurality of road surface point clouds; performing three-dimensional texture reconstruction on the road surface point cloud to obtain a space curved surface; extracting an index set of apparent texture features from the spatial curved surface; and acquiring friction system data through a road surface friction resistance test vehicle, and establishing a mapping relation between the index set and the friction coefficient. The detection method realizes the three-dimensional reconstruction of the microcosmic texture and the macroscopic texture of the pavement by two non-contact detection methods of laser structure depth meter detection and visible light camera, further establishes the mapping relation between the friction coefficient value acquired by the traditional pavement friction resistance test vehicle and an index set, and can directly obtain the friction coefficient of the pavement by the index set and the mapping relation obtained by the method.
Description
Technical Field
The invention relates to the detection of friction coefficient of a road surface, in particular to a non-contact detection method of the friction coefficient of the road surface.
Background
The anti-skid capability of the road surface is not determined unilaterally by the tire or the road condition, but by the interaction between the two, and factors affecting the anti-skid of the road surface include: the characteristics of the road surface (texture), the characteristics of the contact surface (dryness/wetness, whether the road surface is polluted or not), the tire state (speed, pattern, tire pressure and slip ratio) and the like, wherein in the texture of the road surface, the friction coefficient of the road surface in a dry state is mainly influenced by the micro texture (1 mu m-1 mm), and the friction coefficient of the wet road surface is influenced by the macro texture (1 mm-10 mm). How to accurately determine the friction coefficient of the road surface according to the factors becomes an urgent technical problem to be solved.
Disclosure of Invention
The invention aims to provide a non-contact type road surface friction coefficient detection method, which realizes three-dimensional reconstruction of micro and macro texture of a road surface by two non-contact type detection methods of laser structure depth instrument detection and visible light camera detection, further establishes a mapping relation between a friction coefficient value acquired by a traditional road surface friction resistance test vehicle and an index set, and can directly obtain the friction coefficient of the road surface through the index set and the mapping relation obtained by the method.
The technical scheme adopted for realizing the aim of the invention is a non-contact road surface friction coefficient detection method, which comprises the following steps:
s1, three-dimensional texture scanning is carried out on a road surface to obtain a plurality of road surface point clouds;
s2, performing three-dimensional texture reconstruction on the road surface point cloud to obtain a space curved surface;
s3, extracting an index set of apparent texture features from the space curved surface;
and S4, acquiring friction system data through a road surface friction resistance test vehicle, and establishing a mapping relation between the index set and the friction coefficient.
In the technical scheme, a laser structure depth meter is used for scanning the road surface to obtain a plurality of road surface point clouds with microscopic textures, and a visible light camera is used for scanning the road surface to obtain a plurality of road surface point clouds with macroscopic textures.
In the technical scheme, the road surface point cloud is combined into complete road surface three-dimensional information through a splicing algorithm.
In the technical scheme, the three-dimensional reconstruction algorithm of three-dimensional Delaunay triangulation is adopted to carry out three-dimensional reconstruction on the three-dimensional information of the road surface to obtain a space curved surface.
In the technical scheme, the point cloud data are subjected to feature analysis in the space curved surface, and an index set of a surface normal vector, an MTD (maximum transmission digital) construction depth, a contour single peak average interval and a three-dimensional texture morphology fractal dimension is extracted.
In the technical scheme, the mapping relation between the index set and the friction coefficient is established by cluster analysis and regression analysis of the index set and the friction coefficient data.
Further, the non-contact type road surface friction coefficient detection method further comprises the following steps: and (3) performing rough set analysis to aim at the correlation with the friction coefficient, and adjusting parameters of a clustering analysis algorithm to pursue the minimum sum of squares of fitting errors.
Further, the non-contact type road surface friction coefficient detection method further comprises the following steps: and deeply learning the joint probability distribution of the index set and the friction coefficient through an energy function of the Boltzmann machine, and judging the accuracy of the mapping relation according to the joint probability distribution.
The method comprises the steps that firstly, a moving platform formed by a road surface laser structure depth instrument measuring system and a 2D visible light camera in a non-contact type detecting mode is continuously scanned, point clouds can be converted into the same coordinate system to be corrected, wherein noise can be eliminated through a Kalman filter after microscopic texture information of depth scanning is constructed through the laser, and influence of small gray values on road surface texture structure information can be filtered through fixed threshold filtering by macroscopic texture information scanned by the visible light camera. And then, the corrected data of the two sources need to be combined with the characteristics and the spatial position relation of the point cloud, and the three-dimensional reconstruction is realized after the data is finely registered. And finally, calculating key geometric parameters by using the three-dimensional road surface reconstruction graph to extract an index set of relevant characteristics, thereby establishing a mapping relation between the friction coefficient value acquired by the traditional road surface friction resistance test vehicle and the index set. In the operation of actually detecting the friction system, only the index set is required to be obtained by the method, and then the friction coefficient of the road surface can be calculated through the obtained mapping relation.
Compared with the traditional professional pavement friction resistance test vehicle which needs a professional to measure on the spot and needs professional personnel and instrument equipment, the method only needs a laser structure depth instrument to measure and a visible light camera, the needed equipment is simple and economical, the operation is convenient, professional measuring personnel do not need to be equipped, the expenditure of the traditional measurement is saved, and the accuracy of the detection method is high.
Drawings
FIG. 1 is a flow chart of a non-contact road surface friction coefficient detection method according to the present invention.
FIG. 2 is a diagram of raw displacement data of a camera.
Fig. 3 is a schematic diagram of the displacement data obtained by performing kalman filter correction on fig. 2.
Fig. 4 is a schematic diagram of the point cloud before registration.
Fig. 5 is a schematic diagram of the registered point cloud.
Fig. 6 is a schematic diagram of the Delaunay triangulation process.
Fig. 7 is a schematic diagram of a three-dimensional reconstruction model.
FIG. 8 is a schematic diagram of a surface normal vector.
FIG. 9 is a schematic representation of the profile unimodal average spacing.
Detailed Description
The invention is described in further detail below with reference to the figures and the specific embodiments.
As shown in fig. 1, the non-contact road surface friction coefficient detection method of the present invention includes:
s1, three-dimensional texture scanning is carried out on the road surface.
The invention obtains a plurality of road surface point clouds with microscopic textures by scanning a road surface through a laser structure depth meter and obtains a plurality of road surface point clouds with macroscopic textures by scanning the road surface through a visible light camera, and the two scanning modes are respectively explained as follows:
the road surface is scanned by the laser structure depth instrument, the laser structure depth instrument is an instrument for measuring the surface of material particles and the depth change condition among the material particles by utilizing the principle of laser ranging, and the vehicle-mounted laser structure depth instrument is an intelligent instrument and can be used for measuring the road surface structure depth under the normal driving conditions of no serious damage and disease, no accumulated water, accumulated snow, mud and the like. At present, a laser structure depth instrument on the market is relatively mature, the resolution ratio of measurement can reach 0.02mm, the measurement of the apparent micro texture of a pavement can be met, and the noise is eliminated through a Kalman filter after the micro texture information of depth scanning is constructed by laser.
The visible light camera is used for scanning the road surface, the visible light camera is fixed when the macroscopic texture of the road surface is scanned, the plane of the camera is approximately parallel to the surface, and the distance between the camera and the plate is basically consistent. In consideration of the accuracy of the camera, the camera is generally set at a height of 50 to 80cm from the road surface. Generally speaking, in the case of selecting a visible light camera and fixing the distance, if the quality of the acquired information of the surface is high, the scanning range is small, and this time can be solved by moving the camera to continuously scan and then registering the point cloud in a later period. In order to meet the requirement of outdoor mobile acquisition of three-dimensional information of the road surface, the visible light camera can be integrated through the embedded development board. The size and power consumption of the scanning equipment can be reduced by adopting the embedded development board, and the program developed based on the embedded development board can have better consistency and performance. The embedded development board has single function and simple and consistent system architecture, and is very suitable for serving as edge computing equipment for three-dimensional scanning of a road surface, and macroscopic texture information scanned by a visible light camera filters the influence of small gray level values on texture structure information of the road surface through fixed threshold filtering.
And S2, performing three-dimensional texture reconstruction on the road surface point cloud obtained in the S1 to obtain a space curved surface.
And S2.1, forming complete road surface three-dimensional information by the road surface point cloud through a splicing algorithm.
Because the scanning range is limited each time, a plurality of point clouds obtained at different positions need to be spliced to form complete three-dimensional information. The point clouds obtained in each frame are mutually independent, and the three-dimensional coordinate values of the point clouds are calculated relative to the coordinate system at the moment. Therefore, the point cloud coordinates of each frame need to be converted into a world coordinate system according to the spatial position, and the point cloud can be spliced. By taking the macro texture collected by the camera as an example, the change of the pose of the camera in the shooting process can be calculated according to the change of each frame of image in the video. During shooting, the first frame is taken as a world coordinate system as a default, and the displacement of each subsequent frame is the displacement of the camera coordinate system relative to the world coordinate system at the moment.
Translating vector t = [ t ] of camera of each frame 1 ,t 2 ,t 3 ] T And Euler angle vector o = [ r, p, y] T And (3) converting the original displacement data into a world coordinate system, and drawing a displacement curve of the camera, wherein the original displacement data has more abrupt abnormal points, and cannot be directly used for coordinate conversion of the point cloud data as shown in fig. 2.
The camera pose satisfies markov, i.e. the pose at time k is only correlated with the pose at time k-1, but not with the previous pose. In order to eliminate the influence of noise in the observation process, kalman filtering can be adopted to carry out optimal estimation on the motion state. Kalman filtering is an algorithm for optimally estimating the state of a system by using a linear system state equation and combining system observation data. The result of the optimization of the camera motion trajectory using kalman filtering is shown in fig. 3. It can be seen that kalman filtering performs a good filtering action on noise.
And converting the point cloud into the same coordinate system by using the corrected displacement data. Although the kalman filter can better eliminate noise, a slight displacement deviation still exists in the corrected point cloud, and as shown in fig. 4, the point cloud needs to be finely registered in combination with the characteristics of the point cloud.
The ICP algorithm directly selects two points with the closest distance from the point clouds to be matched through linear search as corresponding points, is a violent exhaustive algorithm, has high dependency on initial displacement of the point clouds, and is easy to cause the situation of non-convergence when the initial distance of the two point clouds is far. Because the feature of the plate point cloud is single, a large registration error can also occur when two original point clouds are directly registered. This is also the reason why coordinate transformation of the original point cloud is necessary before ICP algorithm, and after ICP algorithm point cloud registration, it is shown in fig. 5.
And S2.2, performing three-dimensional reconstruction on the three-dimensional information of the road surface by adopting a three-dimensional Delaunay triangulation curved surface reconstruction algorithm to obtain a space curved surface.
As shown in fig. 6, the step of using three-dimensional Delaunay triangulation in this embodiment includes:
(1) Randomly selecting 4 non-coplanar points from the point cloud set P;
(2) And (3) arbitrarily taking a new point Pi epsilon P, and judging the position relation between Pi and the tetrahedron formed by the 4 points in the step (1):
a) If Pi is outside the tetrahedron, calculating the distances from Pi to four surfaces of the tetrahedron, and selecting the surface closest to Pi to form a new tetrahedron with Pi, as shown in FIG. 6 (a);
b) If Pi is inside the tetrahedron, dividing Pi and 4 surfaces of the tetrahedron into new 4 tetrahedrons, as shown in FIG. 6 (b);
c) If Pi is on one surface of the tetrahedron, the triangular plate where Pi is located is divided into 3 new triangular plates, and the tetrahedron is divided into 3 new tetrahedrons with the other vertex out of the surface, as shown in FIG. 6 (c);
d) If Pi is on one edge of the tetrahedron, dividing the tetrahedron into 2 new tetrahedrons according to the edge where Pi is located, as shown in FIG. 6 (d);
(3) And (3) checking whether isolated points exist, if so, continuing the step (2), otherwise, ending, and displaying the three-dimensional reconstructed space curved surface model as shown in fig. 7.
S3, extracting apparent texture features from the space curved surface;
and carrying out feature analysis on the point cloud data in the space curved surface, and extracting an index set of a surface normal vector, an MTD (maximum-mean-distance) construction depth, a contour single-peak average interval and a three-dimensional texture morphology fractal dimension.
The normal vector of the surface is an important attribute of the surface of the road surface, the three-dimensional reconstruction of the road surface also needs to realize an accurate three-dimensional reconstruction result by means of the normal vector, the common method for calculating the normal vector is a method based on local surface fitting, and a sampling plane of point cloud is assumed to be smooth everywhere, so that the local neighborhood of any point can be fitted by using a local plane. The normal vector can be solved by analyzing the eigenvalue vector and the principal component of the eigenvalue of the point cloud within the target point domain, as shown in fig. 8.
Namely, a mesh grid of the point cloud surface is constructed according to data of points in the three-dimensional point cloud data and field information of the points, and then the mesh grid calculates a surface normal.
The construction depth (MTD) is one of the important indexes of the existing specifications for representing the skid resistance of the road surface, and the MTD is the arithmetic mean value of the sum of the distances from each point of the contour to the peak line of the contour in the sampling range, and is shown as the following formula:
namely the arithmetic mean value of the height difference from each point in the sampling range to the highest point in the range in the three-dimensional point cloud data.
The two-dimensional profile single peak average spacing S is defined as the average of the single peak spacing (the projected length of the distance between two adjacent single peak peaks on the centerline) of the profile within the sampling length, and it is also necessary to replace the corresponding length average value with the area average value for the three-dimensional space, as shown in fig. 9.
The method is used for calculating the average value of the unimodal intervals in a road table three-dimensional reconstruction model, namely in a unit sampling range, the unimodal interval of the profile is a main evaluation parameter of the transverse information of the surface of the profile, the peak density of the profile is represented, and the method has important significance for evaluating the stability and the wear resistance of the profile.
The specific calculation method of the fractal box dimension of the three-dimensional texture morphology comprises the following steps: firstly, some cubes with the side length of L are constructed and used for covering a three-dimensional profile curved surface, and the number N of boxes which are intersected with the profile curved surface when the cubes with the side length of L correspond to different side lengths is calculated L And for the three-dimensional reconstruction model of the road surface, setting a cubic box with the side length of L on a plane of vertical projection of a three-dimensional space curved surface. I.e. the entire projection plane is divided into an L x L grid. Stacking cubes on each grid until the cubes are just intersected with the curved surface, calculating the number of the cubes on the grid, and finally accumulating the number of all boxes to obtain N L 。
Drawing logN L 1/log L log-log curve, fractal dimension D is the slope of the curve in the log-log coordinate system.
And S4, extracting an index set according to the extracted apparent texture features, acquiring friction system data through a pavement friction resistance test vehicle, and establishing a mapping relation between the index set and the friction coefficient.
In this embodiment, a mapping relationship between the index set and the friction coefficient is established by cluster analysis and regression analysis of the index set and the friction coefficient data.
The coefficient of friction of a runway is related to the material contacting the object, the degree of surface smoothness, the degree of wetting and drying, the surface temperature, the relative speed of movement, etc. And respectively selecting different functional areas of 2-3 asphalt tracks and cement tracks as test scenes. Two measurement experiments were carried out simultaneously under different working conditions: firstly, acquiring the friction coefficient value by adopting the conventional road surface friction resistance test vehicle; and secondly, performing apparent texture feature analysis at the same position by using the method, and extracting index sets such as surface normal vectors, MTD (maximum Transmission data) construction depth, contour unimodal average distance, three-dimensional texture topography fractal dimension and the like. Through python or MATLAB software, clustering analysis and regression analysis are carried out on the index set and the friction coefficient data, and a mapping relation between the index set and the friction coefficient is established, wherein the specific operation process is as follows:
firstly, clustering analysis is adopted for index set data, the clustering analysis is an unsupervised learning process, a classification standard is not required to be given in advance in the classification process, and the clustering analysis can automatically classify the index set data from sample data. And through cluster analysis, acquiring the distribution condition of the index set data, observing the characteristics of each cluster of data, and further analyzing a specific cluster set in a centralized manner.
For each cluster of clustering index set, establishing a multiple regression model, setting a dependent variable Y (actually measured road surface friction coefficient), wherein k independent variables influencing the dependent variable are X respectively 1 ,X 2 ,...,X k (surface normal vector, MTD structure depth, contour unimodal average distance, three-dimensional texture morphology fractal dimension), and assuming that the influence of each independent variable on a dependent variable Y is linear, namely under the condition that other independent variables are not changed, the average value of Y is along with the independent variable X i We have a global regression model, see formula (3):
Y＝β 0 +β 1 X 1 +β 2 X 2 +...+β k X k +ε (3)
wherein, beta 0 ,β 1 ,β 2 ,...,β k Are regression parameters.
Regression analysis has three basic tasks: 1) The model parameters are estimated using the sample data. 2) Hypothesis testing is performed on the model parameters. 3) A regression model is applied to make predictions for the variables (to the face friction coefficient).
Taking expectations for both sides of equation (3), one can obtain:
E(Y|X 1 ,X 2 ,...,X k )＝β 0 +β 1 X 1 +β 2 X 2 +...+β k X k (4)
equation (4) is called the general regression equation, E (Y | X) 1 ,X 2 ,...,X k ) Expressed in a given argument X i The conditional mean of the values Y was observed under the conditions of (1). Wherein beta is 0 ,β 1 ,β 2 ,...,β k Are often unknown and therefore give corresponding estimates of the overall parameter based on sample observations
At this point, the sample regression equation (5) is obtained.
Wherein the content of the first and second substances,
is E (Y | X) 1 ,X 2 ,...,X k ) The point estimate of (2). Then, parameter estimation is carried out through least square estimation, and the following parameters are set:
q is respectively to
The partial derivative is calculated to be equal to 0, resulting in:
solving the equation set to obtain the estimated value of the parameter
Through cluster analysis, the relation which is potentially difficult to observe in the index set data is found out, the index set data with similar characteristics are subjected to classification preprocessing, and a multiple regression model is established based on the classification data, so that rapid conversion from extraction of road surface texture information to estimation of road surface friction coefficient is realized.
In order to ensure the accuracy of model establishment, the embodiment may further perform a rough set analysis, and the correlation r with the friction coefficient is taken as a target, as shown in formulas (8) and (8), and the sum of squares of fitting errors is pursued to be minimum by adjusting parameters of a clustering analysis algorithm.
Wherein n represents the number of samples, x i Represents the i-th sample friction coefficient test value,
represents the average value, y i Represents the estimated value of the friction coefficient of the ith sample,
represents the average thereof.
The rough set algorithm is a mathematical tool for describing incompleteness and uncertainty, can effectively analyze various incomplete information such as inaccuracy and inconsistency, can analyze and reason data, finds implicit knowledge from the incomplete information, reveals potential rules, and can actually quantify the predictive capability of different elements in an index set on the friction coefficient. The main idea is to derive the decision or classification rule of the problem through knowledge reduction on the premise of keeping the classification capability unchanged. For complex problems, the rough set theory is adopted, and more prior knowledge is not needed, and the data information of the decision table is analyzed. In this way, important factors which characterize the friction coefficient can be found out, and irrelevant factors are eliminated.
The rough set analysis mainly aims to adjust parameters of a clustering analysis algorithm by using the minimization of errors of a pre-estimated value and an actual value as a guide, and finally, a regression analysis model is corrected. The model correction takes the actually measured data as a standard, the target is clear, the accuracy of the prediction model can be better improved, and the prediction performance is enhanced.
In order to verify the mapping relationship between the index set obtained by the method and the friction coefficient value acquired by adopting the traditional road surface friction resistance test vehicle, the embodiment also provides a verification scheme, the verification scheme is realized by taking the energy function of the boltzmann machine as a deep learning model, and the method specifically comprises the following steps:
the joint probability distribution is defined using an energy function:
wherein E (X) is Boltzmann machineZ is the guaranteed sigma x P (x) =1 allocation function,
E(x)＝-x T Ux-b T x
x is a three-dimensional binary random vector, U is a weight matrix of three-dimensional model parameters, b is a bias vector
A group of training samples are taken from friction coefficient values acquired by a traditional road surface friction resistance test vehicle and a three-dimensional binary random vector is established by an index set correspondingly acquired by the method, the joint probability distribution of variables is obtained through learning of a Boltzmann machine based on maximum likelihood, finally, the similarity of the joint probability distribution and a function of a mapping relation is compared, if the similarity of the two is in a specified threshold value range, the mapping relation can truly reflect the relation between an actual friction coefficient value and the index set, wherein the threshold value can be determined after the calculation of a plurality of groups of data and the comparison of actual test data, so that the friction coefficient can be calculated through the index set and the mapping relation acquired by the method, and the calculated friction coefficient is close to the actual friction coefficient.
Claims (8)
1. A non-contact type road surface friction coefficient detection method is characterized by comprising the following steps:
s1, three-dimensional texture scanning is carried out on a road surface to obtain a plurality of road surface point clouds;
s2, performing three-dimensional texture reconstruction on the road surface point cloud to obtain a space curved surface;
s3, extracting an index set of apparent texture features from the space curved surface;
and S4, acquiring friction system data through a road surface friction resistance test vehicle, and establishing a mapping relation between the index set and the friction coefficient.
2. The non-contact type road surface friction coefficient detection method according to claim 1, characterized in that: and scanning the road surface by a visible light camera to obtain a plurality of road surface point clouds with macroscopic textures.
3. The non-contact type road surface friction coefficient detection method according to claim 2, characterized in that: and forming the road surface point cloud into complete road surface three-dimensional information through a splicing algorithm.
4. The non-contact type road surface friction coefficient detection method according to claim 3, characterized in that: and performing three-dimensional reconstruction on the three-dimensional information of the road surface by adopting a three-dimensional Delaunay triangulation curved surface reconstruction algorithm to obtain a space curved surface.
5. The non-contact type road surface friction coefficient detection method according to claim 3, characterized in that: and performing characteristic analysis on the point cloud data in the space curved surface, and extracting an index set of a surface normal vector, MTD (maximum transmission data) construction depth, contour single peak average interval and three-dimensional texture morphology fractal dimension.
6. The non-contact type road surface friction coefficient detection method according to claim 5, characterized in that: and performing cluster analysis and regression analysis on the index set and the friction coefficient data to establish a mapping relation between the index set and the friction coefficient.
7. The non-contact type road surface friction coefficient detection method according to claim 5, characterized in that: further comprising: and (3) the correlation with the friction coefficient is taken as a target through rough set analysis, and the fitting error sum of squares is pursued to be minimum through adjusting parameters of a clustering analysis algorithm.
8. The non-contact type road surface friction coefficient detection method according to claim 6 or 7, characterized in that: and deeply learning the joint probability distribution of the index set and the friction coefficient through an energy function of the Boltzmann machine, and judging the accuracy of the mapping relation according to the joint probability distribution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210847074.3A CN115222916A (en) | 2022-07-07 | 2022-07-07 | Non-contact type road surface friction coefficient detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210847074.3A CN115222916A (en) | 2022-07-07 | 2022-07-07 | Non-contact type road surface friction coefficient detection method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115222916A true CN115222916A (en) | 2022-10-21 |
Family
ID=83611953
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210847074.3A Pending CN115222916A (en) | 2022-07-07 | 2022-07-07 | Non-contact type road surface friction coefficient detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115222916A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117074291A (en) * | 2023-10-17 | 2023-11-17 | 西南交通大学 | Non-contact texture friction prediction method |
-
2022
- 2022-07-07 CN CN202210847074.3A patent/CN115222916A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117074291A (en) * | 2023-10-17 | 2023-11-17 | 西南交通大学 | Non-contact texture friction prediction method |
| CN117074291B (en) * | 2023-10-17 | 2024-01-02 | 西南交通大学 | Non-contact texture friction prediction method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhang et al. | Automatic pavement defect detection using 3D laser profiling technology | |
| CN112465948B (en) | Vehicle-mounted laser pavement point cloud rarefying method capable of retaining spatial features | |
| CN105678734B (en) | A kind of heterologous test image scaling method of image matching system | |
| CN110473187B (en) | Object-oriented line scanning three-dimensional pavement crack extraction method | |
| CN102750449B (en) | Point cloud linear feature extraction method based on substep three-dimensional space and feature dimension mapping | |
| Liu et al. | An improved differential box counting method to measure fractal dimensions for pavement surface skid resistance evaluation | |
| CN115222916A (en) | Non-contact type road surface friction coefficient detection method | |
| CN107288016A (en) | A kind of bituminous paving milling quality evaluating method based on three-dimensional laser detection technique | |
| CN103456029B (en) | The Mean Shift tracking of a kind of anti-Similar color and illumination variation interference | |
| JP2023543453A (en) | Tread depth estimation system and method | |
| CN108931206A (en) | Method for distinguishing is known for rail profile outlier detection and effective profile | |
| CN105205816A (en) | Method for extracting high-resolution SAR image building zone through multi-feature weighted fusion | |
| Tonbul et al. | Semi-automatic building extraction from Worldview-2 imagery using Taguchi optimization | |
| CN113240735B (en) | Slope displacement activity monitoring method | |
| Elunai et al. | Asphalt concrete surfaces macrotexture determination from still images | |
| Chen et al. | 3D pavement data decomposition and texture level evaluation based on step extraction and pavement-transformer | |
| Leclerc et al. | Measuring the self-consistency of stereo algorithms | |
| CN117036326A (en) | Defect detection method based on multi-mode fusion | |
| CN116645348A (en) | High-speed loop linear fitting method based on point cloud random sampling | |
| CN112986948B (en) | Building deformation monitoring method and device based on InSAR technology | |
| Liu et al. | Comparison of pavement texture measurements from a three-dimensional profiler and a circular track meter at MnROAD test facilities | |
| JP2023545657A (en) | System and method for determining tread depth | |
| CN112927204B (en) | Pavement water seepage performance evaluation method based on key water seepage point identification | |
| CN112859073B (en) | Road damage assessment method based on PSInSAR technology | |
| CN117152149B (en) | Edge smoothness detection method and system for wafer dicing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |