A method for evaluating the three-dimensional evenness of asphalt pavement based on two-dimensional power spectral density

Abstract

The present invention discloses a method for evaluating the three-dimensional smoothness of asphalt pavement based on two-dimensional power spectral density, which relates to the technical field of road detection, and establishes the connection between the evaluation method of three-dimensional asphalt pavement smoothness and the current smoothness evaluation standard. The key technical points are to calculate the two-dimensional power spectral density of the three-dimensional asphalt pavement and then convert it into the longitudinal average pavement power spectral density to realize the evaluation of the three-dimensional asphalt pavement smoothness. Compared with the traditional method of evaluating smoothness by calculating the power spectral density of a certain longitudinal section, by adding transverse section data, not only is the evaluation of the three-dimensional asphalt pavement smoothness more comprehensive, but also the fluctuation of the spectral line is reduced, and the evaluation accuracy is improved, which has strong practicability.

Description

Technical Field

[0001] The present application relates to the technical field of road detection, and particularly to a method for evaluating the evenness of a three-dimensional asphalt pavement based on two-dimensional power spectral density. Background Art

[0002] Evenness is one of the main technical indicators for evaluating the quality of asphalt pavements and ensuring driving safety and comfort. Research results show that the deterioration of pavement quality not only leads to a large number of car crashes, but also the unevenness of the road surface affects the driving speed, increases the rolling resistance of the vehicle, accelerates the wear of automotive parts, and increases the maintenance cost of the vehicle; in addition, the uneven road surface will stimulate the vehicle to generate vertical dynamic loads, and this vertical impact load will also be reacted on the road surface, exacerbating the occurrence of road diseases.

[0003] Generally, according to the detection principle, the evenness detection technology can be divided into two types: response type and profile type, mainly including the three-meter straightedge method, continuous evenness meter, hand-pushed evenness meter, vehicle-mounted bump integrator, vehicle-mounted laser evenness meter, etc. The above detection instruments can usually only obtain two-dimensional elevation data for the detection of asphalt pavements, and the calculated evenness is usually for a certain longitudinal section, without considering the influence of the transverse direction of the asphalt pavement.

[0004] With the development of information technology, machine vision technology has been widely used in fields such as robots and automation, and has the advantages of rich information, good timeliness, small size, and low energy consumption. Technologies such as RGB-D sensors and binocular vision technology have been used to obtain more comprehensive three-dimensional information of asphalt pavements. Compared with traditional detection means, the new detection technology can not only obtain the longitudinal elevation data of the asphalt pavement, but also obtain the transverse elevation data of the asphalt pavement.

[0005] However, in the prior art, since the calculation method of pavement evenness is mainly for the longitudinal direction and does not consider the transverse influence, the obtained evenness result cannot fully reflect the condition of the road surface. Summary of the Invention

[0006] The present application provides a method for evaluating the evenness of a three-dimensional asphalt pavement based on two-dimensional power spectral density, and its technical purpose is to establish a connection between the evaluation method of the three-dimensional asphalt pavement evenness and the current evenness evaluation standard, so as to more comprehensively evaluate the evenness of the three-dimensional asphalt pavement.

[0007] The above technical purpose of the present application is achieved through the following technical solutions:

[0008] A method for evaluating the evenness of a three-dimensional asphalt pavement based on two-dimensional power spectral density, comprising:

[0009] Step 1: Discretize the collected asphalt pavement elevation data, that is, sample in the spatial regions in the x and y directions at sampling intervals Δx and Δy respectively to obtain a spatial surface with a lattice size of M×N, and determine the spatial frequencies fx and fy in the x and y directions. p and fy q ;

[0010] Step 2: Preprocess the three-dimensional asphalt pavement elevation data using a Hanning window.

[0011] Step 3: Calculate the two-dimensional discrete Fourier transform of the discretized asphalt pavement elevation data as Z(fx p , fy q );

[0012] Step 4: Calculate the two-dimensional power spectral density as PSD(fx p , fy q ) according to Parseval's theorem.

[0013] Step 5: Convert the two-dimensional power spectral density PSD(fx p , fy q ) into a one-dimensional power spectral density PSD(fx p ) for evaluating the longitudinal pavement.

[0014] Furthermore, in Step 1, the discretization process of the collected asphalt pavement elevation data includes:

[0015] The spatial frequency in the x direction of the spatial surface is discretized as:

[0016] fx p = p / (MΔx), p = 0, 1, 2..., M - 1;

[0017] The spatial frequency in the y direction of the spatial surface is discretized as:

[0018] fy q = q / (NΔy), q = 0, 1, 2..., N - 1.

[0019] Furthermore, in Step 2, the preprocessing of the discretized asphalt pavement elevation data using a Hanning window includes:

[0020] The two-dimensional Hanning window is defined (with window energy normalized) as follows:

[0021]

[0022] where (0 ≤ m ≤ M - 1, 0 ≤ n ≤ N - 1).

[0023] Furthermore, in Step 3, the two-dimensional discrete Fourier transform of the discretized asphalt pavement elevation data is calculated as Z(fx p, f q ), including:

[0024]

[0025] where (0 ≤ m ≤ M - 1, 0 ≤ n ≤ N - 1).

[0026] Further, in the said step 4, the two - dimensional power spectral density is calculated according to Parseval's theorem as PSD(f p , f q ), including:

[0027]

[0028] where (0 ≤ m ≤ M - 1, 0 ≤ n ≤ N - 1).

[0029] Further, in the said step 5, the two - dimensional power spectral density PSD(f p , f q ) is converted into a one - dimensional power spectral density PSD(f p ) for evaluating the longitudinal road surface, including:

[0030]

[0031] where (0 ≤ m ≤ M - 1, 0 ≤ n ≤ N - 1).

[0032] Compared with the prior art, the present invention has the following beneficial technical effects:

[0033] In the present invention, through steps 3 - 4, the two - dimensional discrete Fourier transform of the discrete asphalt pavement elevation data is calculated and the two - dimensional power spectral density is calculated according to Parseval's theorem, and the two - dimensional power spectral density of the three - dimensional asphalt pavement is calculated;

[0034] According to step 5, the two - dimensional power spectral density is converted into a one - dimensional power spectral density for evaluating the longitudinal road surface, thus establishing the connection between the three - dimensional asphalt pavement flatness evaluation method and the current flatness evaluation standard, realizing the evaluation of the three - dimensional asphalt pavement flatness, which is more comprehensive and accurate; compared with the traditional method of evaluating flatness by calculating the power spectral density of a certain longitudinal section, step 1 not only makes the evaluation of the three - dimensional asphalt pavement flatness more comprehensive by adding transverse section data, but also can reduce the fluctuation of the spectral line; the method of the present invention can not only better evaluate the flatness of the three - dimensional asphalt pavement, but also has a very high evaluation accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is the flow chart of the method described in this application;

[0036] Figure 2The three-dimensional elevation map of asphalt pavements from Grade A to Grade D simulated for this application;

[0037] Figure 3 The two-dimensional power spectral density map of asphalt pavements from Grade A to Grade D calculated for this application;

[0038] Figure 4 The comparison chart of the spectral lines calculated by the method of this application and the standard spectrum under the pavements of Grade A to Grade D for this application;

[0039] Figure 5 The comparison chart of the spectral lines calculated by the method of this application and the traditional method under the pavements of Grade A to Grade D for this application. Detailed implementation manners

[0040] The technical solution of this application will be described in detail below with reference to the accompanying drawings.

[0041] This application provides a method for evaluating the evenness of three-dimensional asphalt pavements based on two-dimensional power spectral density. Aiming at the problem that the traditional method cannot comprehensively evaluate the evenness of three-dimensional asphalt pavements, this application uses the two-dimensional power spectral density of three-dimensional asphalt pavements to establish the connection between the evaluation method of three-dimensional asphalt pavement evenness and the current evenness evaluation standard, realizing the evaluation of three-dimensional asphalt pavement evenness, which has the characteristics of being more comprehensive and accurate;

[0042] Example 1:

[0043] A method for evaluating the evenness of three-dimensional asphalt pavements based on two-dimensional power spectral density, and its implementation flowchart is as Figure 1 shown, and its calculation steps are as follows:

[0044] In the three-dimensional asphalt pavement to be evaluated in this application, the simulation formula for its elevation is:

[0045]

[0046] where x is the random longitudinal position of the road surface; y is the random transverse position of the road surface; θ i (x, y) is a random number for any row y on the road surface within [0, 2π].

[0047] Step 1, determine that the lengths of the three-dimensional simulated road surface in the x and y directions are 160 m and 6 m respectively, and take the sampling intervals Δx = 0.1 m and Δy = 0.1 m. Combine the geometric mean of the evenness coefficients of Grade A pavements, which is 16×10 -6 m 3 , to simulate Grade A pavements. Similarly, select the corresponding evenness coefficients for Grades B, C, and D in turn to generate the elevations of three-dimensional asphalt pavements of the corresponding grades.

[0048] Step 2, determine the spatial frequencies f p and f q;

[0049] The spatial frequency in the x - direction of the spatial surface is discretized as:

[0050] f p = p / (MΔx), p = 0, 1, 2..., M - 1;

[0051] The spatial frequency in the y - direction of the spatial surface is discretized as:

[0052] f q = q / (NΔy), q = 0, 1, 2..., N - 1;

[0053] Step 3, pre - process the three - dimensional asphalt pavement elevation data using a Hanning window;

[0054]

[0055] where (0 ≤ m ≤ M - 1, 0 ≤ n ≤ N - 1).

[0056] Step 4, calculate the two - dimensional discrete Fourier transform of the discretized asphalt pavement elevation data as Z(f p , f q );

[0057]

[0058] where (0 ≤ m ≤ M - 1, 0 ≤ n ≤ N - 1).

[0059] Step 5, calculate the two - dimensional power spectral density as PSD(f p , f q ) according to Parseval's theorem;

[0060]

[0061] where (0 ≤ m ≤ M - 1, 0 ≤ n ≤ N - 1).

[0062] Step 6, convert the two - dimensional power spectral density PSD(f p , f q ) to the one - dimensional power spectral density PSD(f p ) for evaluating the longitudinal pavement;

[0063]

[0064] where (0 ≤ m ≤ M - 1, 0 ≤ n ≤ N - 1).

[0065] In this embodiment, as Figure 4 shown, in the spatial frequency interval [0.011, 2.83] m -1Under such circumstances, the calculated power spectral density spectral lines of the three-dimensional asphalt pavement at grades A to D coincide with the standard spectral lines of the corresponding grades, which is feasible for the evaluation of the evenness of the three-dimensional asphalt pavement.

[0066] In this embodiment, as Figure 5 shown, the traditional method is to select the longitudinal section elevation data at the transverse 2m (y = 2m) of each level of simulated pavement and calculate its corresponding power spectral density. Through Figure 5 it can be seen that the power spectral density calculated from the elevation data of a certain longitudinal section in the transverse direction has a large fluctuation range. However, the power spectral density spectral line of the pavement obtained by the method of the present invention coincides with the standard spectral line, and the fluctuation of the spectral line is reduced. Therefore, the method of the present invention can not only better evaluate the evenness of the three-dimensional asphalt pavement, but also has a high evaluation accuracy.

Claims

1. A method for evaluating the flatness of a three-dimensional asphalt pavement based on two-dimensional power spectral density, characterized in that, it includes the following specific steps: S1. Discretize the collected elevation data of the asphalt pavement; S2. Preprocess the three-dimensional elevation data of the asphalt pavement using a Hanning window; S3. Calculate the two-dimensional discrete Fourier transform of the discretized asphalt pavement elevation data as Z(f p , f q ); S4. Calculate the two-dimensional power spectral density as PSD(f p , f q ) according to Parseval's theorem; S5. Convert the two-dimensional power spectral density PSD(f p , f q ) into the one-dimensional power spectral density PSD(f p ) for evaluating the longitudinal road surface.

2. The method for evaluating the flatness of a three-dimensional asphalt pavement based on two-dimensional power spectral density according to claim 1, characterized in that, in S1, the elevation data of the asphalt pavement is obtained by binocular vision or a planar array lidar.

3. The method for evaluating the flatness of a three-dimensional asphalt pavement based on two-dimensional power spectral density according to claim 1, characterized in that, S1 includes the following specific steps: S11, sample in the spatial regions in the x and y directions at sampling intervals Δx and Δy respectively to obtain a spatial surface with a dot matrix size of M×N, and determine the spatial frequencies f p and f q ; S12. The spatial frequency in the x direction of the spatial surface is discretized as: f p = p / (MΔx), p = 0, 1, 2..., M - 1; S13. The spatial frequency in the y direction of the spatial surface is discretized as: f q = q / (NΔy), q = 0, 1, 2..., N - 1.

4. The method for evaluating the flatness of a three-dimensional asphalt pavement based on two-dimensional power spectral density according to claim 1, characterized in that, the preprocessing of the discretized elevation data of the asphalt pavement using a Hanning window in S2 includes: The two-dimensional Hanning window is defined as follows: where (0 ≤ m ≤ M - 1, 0 ≤ n ≤ N - 1).

5. The method for evaluating the flatness of a three-dimensional asphalt pavement based on two-dimensional power spectral density according to claim 1, characterized in that, In S3, the two-dimensional discrete Fourier transform of the discretized asphalt pavement elevation data is calculated as Z(f p , f q ), including: where (0 ≤ m ≤ M - 1, 0 ≤ n ≤ N - 1).

6. The method for evaluating the flatness of a three-dimensional asphalt pavement based on two-dimensional power spectral density according to claim 1, characterized in that, In S4, calculating the two-dimensional power spectral density as PSD(f p , f q ), including: where (0 ≤ m ≤ M - 1, 0 ≤ n ≤ N - 1).

7. The method for evaluating the flatness of a three-dimensional asphalt pavement based on two-dimensional power spectral density according to claim 1, characterized in that, In S5, the two-dimensional power spectral density PSD(f p ,f q ) is converted into the one-dimensional power spectral density PSD(f p ) for evaluating the longitudinal road surface, including: where (0 ≤ m ≤ M - 1, 0 ≤ n ≤ N - 1).

Images