A device and method for rapid detection of moisture content of a pavement asphalt mixture

Abstract

The application relates to the technical field of water content determination processing, in particular to a rapid detection device and method for water content of road asphalt mixture, and specifically comprises the following steps: collecting road electromagnetic wave signals through a ground penetrating radar and dividing windows, and obtaining the wave number energy spectrum of the electromagnetic wave signals in each window; interval division is carried out on the basis of the data fluctuation characteristics, the influence weight of each interval is determined on the basis of the wave number size, the energy values of the wave numbers are weighted and summarized through the influence weight, the comprehensive energy index corresponding to each window is constructed, the best cut-off wave number corresponding to each window is calculated in combination with the confusion degree of energy distribution and sequence randomness in the wave number energy spectrum, filtering is carried out in combination with a filtering algorithm, and the water content of the road asphalt mixture is calculated on the basis of the filtered signals; compared with fixed threshold filtering, the fixed threshold filtering can remove noise from the signals while retaining more useful information, improves the denoising effect of the signals, and improves the accuracy of subsequent water content calculation.

Description

Technical Field

[0001] This application relates to the field of moisture content determination and processing technology, specifically to a rapid detection device and method for the moisture content of asphalt mixtures for road surfaces. Background Technology

[0002] With the diversification of people's travel, the health of roads is extremely important. Roads require regular maintenance and repair after long-term use, typically using asphalt mixtures for in-situ cold recycling. When repairing roads, the moisture content of the asphalt mixture is a key indicator affecting the quality of highway engineering. Excessive moisture content significantly reduces the adhesion between asphalt and aggregates, leading to problems such as peeling and loosening of the pavement, seriously affecting the service life of the road and driving safety.

[0003] After cold recycling repair using asphalt mixtures, relevant data, such as moisture content at various test points on the road and electromagnetic wave signals, are typically collected using a nucleus-free density meter and ground-penetrating radar (GPR) to quickly detect the moisture content of the asphalt mixture. However, when using GPR with high-frequency electromagnetic waves, the aggregate distribution in the repaired asphalt mixture is often uneven, with localized concentrations of fine or coarse aggregates (asphalt segregation), leading to abrupt changes in the dielectric constant of those areas. This results in a chaotic "spotted" reflection on the radar profile, easily mistaken for uneven humidity or changes in porosity, causing abnormal deviations in the obtained dielectric constant. This can lead to inaccuracies in moisture content calculations, resulting in significant errors in the final moisture content determination of the asphalt mixture. Summary of the Invention

[0004] To address the aforementioned technical problems, the purpose of this application is to provide a rapid detection device and method for the moisture content of asphalt mixtures for road surfaces. The specific technical solution adopted is as follows:

[0005] In a first aspect, embodiments of this application provide a method for rapid detection of the moisture content of pavement asphalt mixtures, the method comprising the following steps:

[0006] Collect electromagnetic wave signals from the road in the time-space domain;

[0007] The collected electromagnetic wave signals are divided into windows to obtain the wavenumber energy spectrum of the FK spectrum of the electromagnetic wave signals in each window; the wavenumber intervals of each wavenumber energy spectrum are divided based on the data fluctuation characteristics in each wavenumber energy spectrum; the influence weight of each wavenumber interval is determined based on the wavenumber magnitude in each wavenumber interval; and the comprehensive energy index corresponding to each window is constructed by combining the energy value of each wavenumber.

[0008] The sequence of energy values ​​in each wavenumber energy spectrum is denoted as the energy sequence; based on the wavenumber energy distribution in the wavenumber energy spectrum and the randomness of data changes in the energy sequence, an abnormal interference index corresponding to each window is constructed.

[0009] Based on the comprehensive energy index and the abnormal interference index, the optimal cutoff wave number corresponding to each window is determined, and the collected electromagnetic wave signal is filtered by the filtering algorithm.

[0010] Based on the filtered electromagnetic wave signal and combined with the simulation model for predicting moisture content, the moisture content of the road surface is determined.

[0011] In one embodiment, the process of dividing the wavenumber energy spectrum into wavenumber intervals is as follows:

[0012] Wavenumbers with energy levels lower than their left-hand neighboring wavenumbers but higher than their right-hand neighboring wavenumbers in the energy spectrum are identified as falling wavenumbers. If all falling wavenumbers are adjacent to falling wavenumbers on both sides, then each falling wavenumber is marked. The maximum and minimum wavenumbers among all marked falling wavenumbers are identified. The interval between the maximum and minimum wavenumbers in the energy spectrum is identified as the middle wavenumber interval, the interval with wavenumbers lower than the minimum wavenumber is identified as the low wavenumber interval, and the interval with wavenumbers higher than the maximum wavenumber is identified as the high wavenumber interval.

[0013] In one embodiment, the process of obtaining the influence weight is as follows:

[0014] In each wavenumber energy spectrum, all wavenumber intervals are numbered in descending order of wavenumber. The influence weight of each wavenumber interval is determined by the numbering, and the influence weight of each wavenumber interval is negatively correlated with the numbering of each wavenumber interval.

[0015] In one embodiment, the process of obtaining the comprehensive energy index is as follows:

[0016] The energy of all wavenumbers in the wavenumber energy spectrum corresponding to each window is weighted and summed using the influence weights of the corresponding wavenumber intervals, and then normalized to obtain the comprehensive energy index corresponding to each window.

[0017] In one embodiment, the process of obtaining the abnormal interference index is as follows:

[0018] Calculate the Shannon entropy of the energy values ​​of all wavenumbers in the wavenumber energy spectrum corresponding to each window; calculate the permutation entropy of each energy sequence; determine the abnormal interference index corresponding to each window based on the Shannon entropy and the permutation entropy, wherein the abnormal interference index is positively correlated with the Shannon entropy and the permutation entropy respectively.

[0019] In one embodiment, the abnormal interference index is the normalized value of the product of the Shannon entropy, the permutation entropy, and the standard deviation.

[0020] In one embodiment, the process of obtaining the optimal cutoff wavenumber is as follows:

[0021] The influence coefficients for constructing the cutoff wavenumbers for each window are determined based on the comprehensive energy index and the abnormal interference index. The influence coefficients are negatively correlated with the comprehensive energy index and the abnormal interference index, respectively. The difference between the preset maximum cutoff wavenumber and the preset minimum cutoff wavenumber is calculated. Based on the difference, the influence coefficients, and the minimum cutoff wavenumber, the optimal cutoff wavenumber for each window is determined.

[0022] In one embodiment, the expression for the influence coefficient is:

[0023] In the formula, The influence coefficients constructed for the cutoff wavenumber of the a-th window; This refers to the comprehensive energy index corresponding to the a-th window; This refers to the abnormal interference index corresponding to the a-th window; It is a preset minimum positive number.

[0024] In one embodiment, the expression for the optimal cutoff wavenumber is:

[0025]

[0026] In the formula, This represents the optimal cutoff wavenumber corresponding to the a-th window. , These are the minimum cutoff wavenumber and the maximum cutoff wavenumber, respectively.

[0027] Secondly, embodiments of this application also provide a rapid detection device for the moisture content of asphalt mixtures for road surfaces, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of any of the methods described above.

[0028] The embodiments of this application have at least the following beneficial effects:

[0029] This application uses ground-penetrating radar to detect high-frequency electromagnetic wave signals from roads. The acquired electromagnetic wave signals are windowed to obtain the wavenumber energy spectrum of the FK spectrum within each window. Based on the data fluctuation characteristics in each wavenumber energy spectrum, wavenumber intervals are divided, and the influence weight of each wavenumber interval is determined based on the wavenumber magnitude, highlighting the influence of noise-related frequency bands and improving the feature discrimination capability. The energy values ​​of each wavenumber are weighted and summarized using the influence weights to construct a comprehensive energy index corresponding to each window, forming a quantitative description of the severity of speckle noise in the overall signal. The wavenumber energy spectrum is then analyzed... The disorder of energy distribution and the randomness of the sequence are used to construct abnormal interference indicators for each window, effectively identifying the characteristics of speckle noise generation. The optimal cutoff wavenumber for each window is calculated using the above characteristic indicators, and then filtered using a filtering algorithm. Compared with fixed threshold filtering, this method removes noise from the signal within each window while retaining more useful information, improving the denoising effect and thus enhancing the accuracy of subsequent moisture content calculation. This avoids the problem of uneven aggregate distribution in the pavement after asphalt mixture repair, which easily leads to speckle noise in high-frequency electromagnetic wave signals and affects the detection of asphalt mixture moisture content. Attached Figure Description

[0030] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0031] Figure 1 A flowchart illustrating the steps of a rapid detection method for the moisture content of asphalt mixtures for road surfaces, provided as an embodiment of this application;

[0032] Figure 2 This is a schematic diagram illustrating the process of obtaining abnormal interference indicators. Detailed Implementation

[0033] To further illustrate the technical means and effects adopted by this application to achieve the intended purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a rapid detection device and method for the moisture content of asphalt mixtures for road surfaces proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0034] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0035] The following description, in conjunction with the accompanying drawings, details the specific scheme of the rapid detection device and method for the moisture content of asphalt mixtures for road surfaces provided in this application.

[0036] Please see Figure 1 The diagram illustrates a flowchart of a rapid detection method for the moisture content of asphalt mixtures for road surfaces, provided in one embodiment of this application. The method includes the following steps:

[0037] Step S1: Collect electromagnetic wave signals in the time-space domain of the road.

[0038] When rapidly testing the moisture content of asphalt mixtures used for in-situ cold recycling, a measurement line is first planned along the lane direction on the road surface to be tested. K key locations are selected as key testing points on the measurement line; in this embodiment, K is set to 200 to ensure sufficient data acquisition. In other embodiments of this application, the implementer can set the value of K according to the actual situation.

[0039] A ground-penetrating radar (GPR) is mounted on a detection vehicle. The GPR transmits high-frequency electromagnetic wave signals and receives reflected signals via an antenna. Since the asphalt mixture used in in-situ cold recycling technology is typically shallow, the antenna center frequency used in this embodiment is 2 GHz. In other embodiments of this application, the implementer can set the frequency of the transmitted electromagnetic waves according to actual conditions.

[0040] After static preheating, driving tests are performed. During driving, reflected electromagnetic wave signals are received in real time via an antenna. To ensure effective capture of high-frequency signals, the sampling frequency of the reflected signals is set to 12GHz in this embodiment, and road location information is recorded synchronously via GPS, thereby obtaining the electromagnetic wave signals of the road in the time-space domain. In other embodiments of this application, the implementer can set the sampling frequency according to the actual situation.

[0041] Meanwhile, the moisture content at each key detection point is collected using a nucleus-free density meter. The device measures the moisture content at each key detection point by emitting electromagnetic wave signals.

[0042] Step S2: The collected electromagnetic wave signal is divided into windows to obtain the wavenumber energy spectrum of the FK spectrum of the electromagnetic wave signal in each window; the wavenumber intervals of each wavenumber energy spectrum are divided based on the data fluctuation characteristics in each wavenumber energy spectrum; the influence weight of each wavenumber interval is determined based on the wavenumber magnitude in each wavenumber interval, and the comprehensive energy index corresponding to each window is constructed by combining the energy value of each wavenumber.

[0043] Considering that there is usually uneven aggregate distribution in the road surface after asphalt mixture repair, and the concentration of fine or coarse aggregate in local areas, i.e. asphalt pavement segregation, it will cause a sudden change in the dielectric constant of the area. The collected electromagnetic wave signal of the road is prone to spot noise, which in turn leads to abnormal detection when calculating the moisture content of the asphalt mixture.

[0044] Based on the above analysis, this application first performs filtering and noise reduction on the high-frequency electromagnetic wave signals received in real time using ground-penetrating radar. The FK filtering algorithm is used to filter the collected high-frequency electromagnetic wave signals from the road. Specifically, the FK filtering algorithm uses a two-dimensional fast Fourier transform to convert the data to the frequency-wavenumber (FK) domain, multiplies it by a filtering window in the FK domain, and then performs an inverse transform back to the time-range domain. However, the speckle noise characteristics vary across different roads during the filtering process; therefore, the FK filtering algorithm needs to be improved to enhance the filtering effect.

[0045] Considering that the unevenness of the road surface varies spatially when using ground-penetrating radar to acquire high-frequency electromagnetic wave signals, the signals acquired during the journey are overlapped and windowed with an overlap rate of 30% to ensure effective edge stitching during subsequent signal synthesis. The signal within each window is analyzed separately. The length of each window is set to 100ms; if a window is shorter than 100ms, mean filling is used to supplement the window. A two-dimensional Fourier transform is performed on the signal within each window to convert it from the time-space domain to the frequency-wavenumber (FK) domain, obtaining the frequency-wavenumber (FK) spectrum corresponding to each window. The spectral characteristics of each window are then analyzed and processed. Overlapping windowing, mean filling, and two-dimensional Fourier transform are all well-known techniques, and their specific processes will not be elaborated further. In other embodiments of this application, the implementer can set the overlap rate and window length according to actual conditions.

[0046] Considering that the window primarily contains effective layered reflection signals, which typically exhibit energy concentration in the low-frequency, low-wavenumber region and good coherence, a higher cutoff wavenumber is required to ensure the preservation of the effective signal. Conversely, when the window is dominated by disordered spot noise caused by aggregate segregation, it displays a dispersed energy distribution across the entire wavenumber range (especially in the high-wavenumber region), necessitating a lower cutoff wavenumber for more powerful noise filtering. Therefore, analysis and processing are performed based on these characteristics.

[0047] Considering that speckle noise generates significant energy in the high wavenumber region of the FK domain, while the energy of a normal, finite signal is concentrated in the low wavenumber region, an analysis is performed to determine if the signal has severe noise. Taking the a-th window as an example, the square of the modulus of the FK spectrum result within that window is used to obtain the corresponding wavenumber energy spectrum. The overall pattern shows that the energy is relatively flat in the low wavenumber range, gradually decreasing in the mid-wavenumber range, and low in the high wavenumber range, but with chaotic fluctuations due to noise. Therefore, wavenumbers with energy less than their left-hand adjacent wavenumber but greater than their right-hand adjacent wavenumber are identified as falling wavenumbers. If both the left and right adjacent wavenumbers of each falling wavenumber are falling wavenumbers, then each falling wavenumber is marked. The maximum and minimum wavenumbers among all marked falling wavenumbers are obtained. The interval between the maximum and minimum wavenumbers in each wavenumber energy spectrum is designated as the mid-wavenumber range, the interval less than the minimum wavenumber is designated as the low-wavenumber range, and the interval greater than the maximum wavenumber is designated as the high-wavenumber range. If no interval that meets the conditions can be detected during the above interval division process, the intervals are divided according to a preset ratio. In this embodiment, the intervals are divided according to the preset ratio as follows: the first 30% of the current wavenumber range is the low wavenumber interval, 30%-70% is the medium wavenumber interval, and the last 30% is the high wavenumber interval.

[0048] The influence weight of each wavenumber interval is determined by the wavenumber magnitude within each wavenumber interval of the wavenumber energy spectrum. The larger the wavenumber in each wavenumber interval of the wavenumber energy spectrum, the greater the influence weight is assigned. Preferably, in this embodiment, the process of obtaining the influence weight is as follows: the three intervals of the wavenumber energy spectrum corresponding to each window are numbered in descending order of wavenumber, so that the high wavenumber interval is interval 1, the medium wavenumber interval is interval 2, and the low wavenumber interval is interval 3. The expression for the influence weight is:

[0049]

[0050] In the formula, This represents the influence weight of the k-th wavenumber interval corresponding to the a-th window; This represents the number of the k-th wavenumber interval corresponding to the a-th window; The table contains the total number of wavenumber intervals corresponding to the a-th window; This represents an exponential function with the natural constant as its base. Using this function, higher weights can be assigned to intervals closer to higher wavenumbers through non-linear weighting.

[0051] In other embodiments of this application, the process of obtaining the influence weight may also be as follows: numbering the three intervals of the wavenumber energy spectrum in order of wavenumber from low to high, and using the number of each wavenumber interval as the influence weight of each wavenumber interval.

[0052] This application quantifies wavenumbers by dividing them into regions and assigns high weights to high wavenumbers through weighted analysis.

[0053] Furthermore, the energies of all wavenumbers in the wavenumber energy spectrum corresponding to each window are weighted and summed using the influence weights to obtain the comprehensive energy index corresponding to each window. Preferably, in this embodiment, the process of obtaining the comprehensive energy index is as follows: calculate the sum of all wavenumber energies within each wavenumber interval, denoted as the first energy of each wavenumber interval; use the influence weights of each wavenumber interval as the weights of the first energy, and calculate the weighted average of the first energy of all wavenumber intervals corresponding to each window; divide the weighted average by the sum of all wavenumber energies in the wavenumber energy spectrum corresponding to each window to achieve normalization, thereby using this ratio as the comprehensive energy index corresponding to that window. The weighted average is a known technique, and the specific process will not be elaborated further.

[0054] In other embodiments of this application, the process of obtaining the comprehensive energy index may also be as follows: calculating the product of the energy of each wavenumber and the influence weight of its corresponding wavenumber interval, and using the normalized value of the sum of the products of all wavenumbers corresponding to each window as the comprehensive energy index corresponding to each window. In this embodiment, the sum of the products of all windows is normalized using the maximum-minimum normalization method. Implementers may also use other normalization functions to normalize the sum of the products, and this application does not impose specific limitations.

[0055] The more severe the speckle noise in the window, the more pronounced the energy in the high-wavenumber region, and the stronger the high-wavenumber energy obtained. After nonlinear weighting, the weight of the high-wavenumber energy is increased, resulting in a larger overall energy index value. When the window is dominated by effective signals, the corresponding high-wavenumber region has lower energy, while the low-wavenumber region has higher energy. After nonlinear weighting, the weight of the low-wavenumber energy is reduced, resulting in a smaller overall energy index value.

[0056] Step S3: Record the sequence of energy values ​​in each wavenumber energy spectrum as an energy sequence; based on the degree of disorder in the wavenumber energy distribution and peak energy distribution in the wavenumber energy spectrum, and the degree of randomness in the data changes in the energy sequence, construct the abnormal interference index corresponding to each window.

[0057] Existing technologies simply calculate the disorder of all energy values ​​in the wavenumber energy spectrum, such as by calculating Shannon entropy, to obtain the feature distribution within the target window. This approach fails to consider the energy fluctuations, random peaks, and varying magnitudes of these fluctuations within the local window after speckle noise occurs. Directly calculating the disorder of energy values ​​in such scenarios is ineffective in capturing the characteristics generated by speckle noise, leading to errors in subsequent filtering and denoising. This solution analyzes the overall anomaly within the signal target window by combining the transformation characteristics of local peaks, thereby obtaining the signal's characteristics within the target window.

[0058] Specifically, considering that the energy distribution of conventional useful signals in the wavenumber energy spectrum corresponding to a window is regular, with energy concentrated in a few specific frequency ranges, while speckle noise appears randomly, its energy distribution range on the frequency axis is scattered and of random size, therefore, taking the a-th window as an example, the degree of disorder in the energy value distribution of all wavenumbers in the wavenumber energy spectrum corresponding to that window is first calculated. The degree of disorder can be variance, standard deviation, information entropy, etc. In this embodiment, the degree of disorder in the energy values ​​of all wavenumbers is the Shannon entropy of the energy values ​​of all wavenumbers. When the speckle noise in the window is severe, the energy situation in the window becomes more disordered due to the random changes in noise, resulting in a larger entropy value. When the window mainly contains effective signals, the energy distribution in the window is relatively regular. If the energy value range in the window is divided into multiple energy value ranges, the energy is mainly distributed in a few energy ranges, resulting in a smaller entropy value of the energy values.

[0059] Furthermore, considering that the energy distribution of a conventional useful signal in the wavenumber energy spectrum corresponding to a window is continuous and regular, but when speckle noise is severe, the mid-to-high wavenumber regions of the wavenumber energy spectrum experience random high-energy abrupt changes due to noise, leading to random jumps in the energy arrangement on the wavenumber. Therefore, the sequence of energy values ​​of all wavenumbers in the wavenumber energy spectrum corresponding to each window is taken as the energy sequence of that window; the permutation entropy of this energy sequence is calculated. In this embodiment, the embedding dimension is set to 3 and the delay time is 1 when calculating the permutation entropy. In other embodiments of this application, the implementer can set the embedding dimension and delay time according to the actual situation.

[0060] It should be noted that, for assessing the randomness of data changes in the energy sequence, implementers may also use other algorithms or indicators, such as approximate entropy, autocorrelation coefficient, etc., to assess the randomness of data changes in the energy sequence. This application does not impose specific restrictions.

[0061] Based on the above analysis, an abnormal interference index is constructed for each window to characterize the correlation and regularity of wavenumbers within each window being abnormally interfered with by speckle noise. Preferably, in this embodiment, the expression of the abnormal interference index is:

[0062]

[0063] In the formula, This refers to the abnormal interference index corresponding to the a-th window; The disorder of the energy distribution in the wavenumber energy spectrum corresponding to the a-th window; Let represent the degree of randomness in the data variation within the energy sequence corresponding to the a-th window. The product is the first product. In this embodiment, the normalized value of the first product of each window is obtained by normalizing the maximum and minimum values ​​of the first product of all windows. Implementers may also use other methods for normalization, and this application does not impose specific restrictions.

[0064] In other embodiments of this application, the expression for the abnormal interference index may also be: ,in, , They are respectively and The normalized value is obtained by using the same normalization method as the first product.

[0065] When the speckle noise within the window is severe, the energy distribution within the window becomes increasingly chaotic due to the random changes in noise, and the peak values ​​caused by the noise fluctuate significantly, resulting in... The larger the value, the more regular the energy distribution within the window becomes when the window mainly contains valid signals. Peak values ​​caused by normal Gaussian noise are more stable, resulting in higher values. The value is relatively small. The smaller the value, the more severe the effect of speckle noise within the window, resulting in lower values. The larger the value, the more severe the abnormal interference. The value increases accordingly. As interference decreases, the acquired... The value then decreases.

[0066] Step S4: Based on the comprehensive energy index and the abnormal interference index, determine the optimal cutoff wave number corresponding to each window, and filter the collected electromagnetic wave signal using a filtering algorithm.

[0067] The influence coefficients for constructing the cutoff wavenumbers of each window are determined based on the comprehensive energy index and the anomalous interference index. The larger the comprehensive energy index and the anomalous interference index, the smaller the influence coefficient. Preferably, in this embodiment, the expression for the influence coefficient is:

[0068]

[0069] In the formula, The influence coefficients constructed for the cutoff wavenumber of the a-th window; This refers to the comprehensive energy index corresponding to the a-th window; This refers to the abnormal interference index corresponding to the a-th window; The value is a preset minimum positive number, used to avoid a denominator of 0. In this embodiment, it is... The value is set to 0.001. In other embodiments of this application, the implementer may set the value according to the actual situation. The value obtained is determined by the severity of noise in the window. The smaller the value, the more... The smaller the value, the less noise is in the window, and the better the acquired data. The larger the value, the more you get. The larger the value.

[0070] In other embodiments of this application, the expression for the influence coefficient may also be: .

[0071] First, set the minimum cutoff wavenumber. With maximum cutoff wavenumber Specifically, since noise typically exists in the mid-to-high wavenumber regions of the wavenumber energy spectrum, and the high wavenumber region mainly contains noise, the mid-wavenumber region of the complete original signal is obtained according to the method described above for distinguishing the mid-wavenumber region. The minimum wavenumber of the mid-wavenumber interval of the complete signal is then used as the minimum truncation wavenumber. The maximum wavenumber in the middle wavenumber range of the complete signal is taken as the maximum truncated wavenumber. Based on the above analysis, the optimal cutoff wavenumber for each window is constructed, expressed as:

[0072]

[0073] In the formula, This represents the optimal cutoff wavenumber for the a-th window. The more severe the noise in the window, the more... The smaller the value, the better. The smaller the value, the more appropriate the use. FK filtering can effectively remove noise; the less noise in the window, the better. The larger the value, the more... The larger the value, the more effective signal is retained during filtering.

[0074] In existing technologies, the cutoff wavenumber is typically set by setting upper and lower limits for wavenumber regions with noise. However, this method does not consider the signal energy distribution and internal feature transformation during actual processing, resulting in a large signal error for water content detection. In this application, the energy distribution and internal feature transformation of the signal within a local time range are fully incorporated, and the cutoff wavenumber is adaptively adjusted to obtain the optimal cutoff wavenumber for the target window, thereby reducing signal error.

[0075] After calculating the optimal cutoff wavenumber for each window, FK filtering is used to filter the high-frequency electromagnetic wave signal within each window, and the filtered window signal is output. FK filtering is a well-known technique, and the specific operation steps will not be described in detail. The processed signals are synthesized by combining the overlapping parts in sequence. The signals of adjacent windows are synthesized by calculating the mean of the overlapping parts, thus completing the filtering process for high-frequency electromagnetic waves.

[0076] Step S5: Based on the filtered electromagnetic wave signal and combined with the simulation model of predicted moisture content, determine the moisture content of the road surface.

[0077] The high-frequency electromagnetic wave signal of the road after the above filtering process was obtained, as well as the moisture content of each key detection point on the road obtained by the nuclear density meter.

[0078] Using gprMax simulation software based on the finite-difference time-domain method, static ground-penetrating radar reflection signals were numerically simulated on cold recycled pavement models with different preset moisture contents to obtain the amplitude of reflected waves and incident waves on the pavement surface, and then the dielectric constant was calculated. In this embodiment, the different preset moisture contents are specifically 2%, 4%, 6%, 8%, and 10%. By fitting the simulation data, the key parameters in the simulation model and the reference dielectric constant (i.e., the dielectric constant under dry conditions) were determined, thereby establishing a preliminary dielectric constant-moisture content model as a theoretical model for predicting road moisture content. In the actual measurement stage, the operation process includes two parts: on-site calibration and continuous detection. For the pavement to be tested, a measurement line was laid out along the lane direction, and a series of key sampling points were selected on the measurement line. At one of the key sampling points, the actual average reflected wave amplitude and moisture content of the key sampling point were obtained by on-site measurement with static ground-penetrating radar and sampling with a nucleus-free density meter, respectively. Based on this, a correction coefficient was calculated to correct the influence of material additives (such as emulsifiers and regenerators), and the theoretical model was calibrated on-site. The processes of constructing the theoretical model, calculating the correction coefficients, and calibrating the theoretical model are all well-known.

[0079] Subsequently, the array ground-penetrating radar is mounted on a detection vehicle and travels at a constant speed along the road measurement line to continuously scan the entire road width. The collected electromagnetic wave signals from the road are processed through the aforementioned filtering steps. The filtered signal data is then input into the corrected dielectric constant-moisture content model, which outputs the moisture content of the asphalt mixture in the road surface. Combined with the road volume, the overall moisture content of the asphalt mixture used in the road surface is obtained. The process of calculating the moisture content using the dielectric constant-moisture content model is well-known and will not be elaborated further.

[0080] A schematic diagram illustrating the process of obtaining abnormal interference indicators is shown below. Figure 2 As shown.

[0081] Based on the same inventive concept as the above method, this application embodiment also provides a rapid detection device for the moisture content of pavement asphalt mixture, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-described rapid detection methods for the moisture content of pavement asphalt mixture.

[0082] In summary, this application provides a rapid method for detecting the moisture content of asphalt mixtures for road surfaces. It utilizes ground-penetrating radar to detect high-frequency electromagnetic signals from the road. The acquired electromagnetic signals are windowed to obtain the wavenumber energy spectrum of the FK spectrum within each window. Based on the data fluctuation characteristics in each wavenumber energy spectrum, wavenumber intervals are divided, and the influence weight of each wavenumber interval is determined based on the wavenumber magnitude. This highlights the influence of noise-related frequency bands and improves the discriminative ability of features. The energy values ​​of each wavenumber are weighted and summarized using these influence weights to construct a comprehensive energy index corresponding to each window, forming a comprehensive energy index for detecting severe speckle noise in the overall signal. The system provides a quantitative description of the situation; it constructs abnormal interference indices for each window based on the disorder and randomness of the energy distribution in the wavenumber energy spectrum, effectively identifying the characteristics of speckle noise generation; it calculates the optimal cutoff wavenumber for each window using the aforementioned indices, and combines this with a filtering algorithm for filtering. Compared to fixed threshold filtering, this method removes noise from the signal within each window while retaining more useful information, improving the denoising effect and thus enhancing the accuracy of subsequent moisture content calculation. This approach avoids the problem of uneven aggregate distribution in asphalt mixture repaired pavements, which can easily lead to speckle noise in high-frequency electromagnetic signals and affect the detection of asphalt mixture moisture content.

[0083] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this application. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0084] The various embodiments in this application are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0085] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the principles of this application should be included within the protection scope of this application.

Claims

1. A method for rapid detection of moisture content of a pavement asphalt mixture, characterized in that, The method includes the following steps: Collect electromagnetic wave signals from the road in the time-space domain; The collected electromagnetic wave signals are divided into windows to obtain the wavenumber energy spectrum of the FK spectrum of the electromagnetic wave signals in each window; the wavenumber intervals of each wavenumber energy spectrum are divided based on the data fluctuation characteristics in each wavenumber energy spectrum; the influence weight of each wavenumber interval is determined based on the wavenumber magnitude in each wavenumber interval; and the comprehensive energy index corresponding to each window is constructed by combining the energy value of each wavenumber. The sequence of energy values ​​in each wavenumber energy spectrum is denoted as the energy sequence; based on the wavenumber energy distribution in the wavenumber energy spectrum and the randomness of data changes in the energy sequence, an abnormal interference index corresponding to each window is constructed. Based on the comprehensive energy index and the abnormal interference index, the optimal cutoff wave number corresponding to each window is determined, and the collected electromagnetic wave signal is filtered by the filtering algorithm. Based on the filtered electromagnetic wave signal and combined with the simulation model for predicting moisture content, the moisture content of the road surface is determined. The process for obtaining the optimal cutoff wavenumber is as follows: The influence coefficients for constructing the cutoff wavenumbers for each window are determined based on the comprehensive energy index and the abnormal interference index. The influence coefficients are negatively correlated with the comprehensive energy index and the abnormal interference index, respectively. The difference between the preset maximum cutoff wavenumber and the preset minimum cutoff wavenumber is calculated. Based on the difference, the influence coefficients, and the minimum cutoff wavenumber, the optimal cutoff wavenumber for each window is determined. The expression for the influence coefficient is: In the formula, The influence coefficients constructed for the cutoff wavenumber of the a-th window; This refers to the comprehensive energy index corresponding to the a-th window; This refers to the abnormal interference index corresponding to the a-th window; It is a preset minimum positive number; The expression for the optimal cutoff wavenumber is: In the formula, represents the optimal cutoff wave number corresponding to the a-th window, , are the minimum cutoff wave number and the maximum cutoff wave number, respectively.

2. The method for rapid detection of water content of a pavement asphalt mixture according to claim 1, characterized in that, The process of dividing the wavenumber energy spectrum into wavenumber intervals is as follows: The wavenumbers whose energy is less than the energy of their left-side adjacent wavenumber and greater than the energy of their right-side adjacent wavenumber in the energy spectrum are identified as falling wavenumbers. If both the left and right adjacent wavenumbers of each falling wavenumber are falling wavenumbers, then each falling wavenumber is marked. The maximum and minimum wavenumbers among all marked falling wavenumbers are identified. The interval between the maximum and minimum wavenumbers in the energy spectrum of each wavenumber is identified as the middle wavenumber interval, the interval less than the minimum wavenumber is identified as the low wavenumber interval, and the interval greater than the maximum wavenumber is identified as the high wavenumber interval.

3. The method for rapid detection of moisture content in asphalt mixtures for road surfaces as described in claim 1, characterized in that, The process of obtaining the influence weights is as follows: In each wavenumber energy spectrum, all wavenumber intervals are numbered in descending order of wavenumber. The influence weight of each wavenumber interval is determined by the numbering, and the influence weight of each wavenumber interval is negatively correlated with the numbering of each wavenumber interval.

4. The method for rapid detection of moisture content in asphalt mixtures for road surfaces as described in claim 1, characterized in that, The process of obtaining the comprehensive energy index is as follows: The energy of all wavenumbers in the wavenumber energy spectrum corresponding to each window is weighted and summed using the influence weights of the corresponding wavenumber intervals, and then normalized to obtain the comprehensive energy index corresponding to each window.

5. The method for rapid detection of moisture content in asphalt mixtures for road surfaces as described in claim 1, characterized in that, The process of obtaining the abnormal interference index is as follows: Calculate the Shannon entropy of the energy values ​​of all wavenumbers in the wavenumber energy spectrum corresponding to each window; calculate the permutation entropy of each energy sequence; determine the abnormal interference index corresponding to each window based on the Shannon entropy and the permutation entropy, wherein the abnormal interference index is positively correlated with the Shannon entropy and the permutation entropy respectively.

6. The method for rapid detection of water content of a pavement asphalt mixture according to claim 5, characterized in that, The abnormal interference index is the normalized value of the product of the Shannon entropy and the permutation entropy.

7. A device for rapid detection of water content of a pavement asphalt mixture, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, The processor implements the steps of the method according to any one of claims 1-6 when executing the computer program.

Images