A device and method for detecting interlayer voids in cement concrete pavement

Abstract

This invention relates to the field of image processing technology, and provides a device and method for detecting interlayer voids in cement concrete pavements. The method includes: acquiring first radar local features and first vibration local features of a first region of the cement concrete pavement; acquiring second radar local features and second vibration local features of a second region; and acquiring third radar local features and third vibration local features of a third region; determining a first local radar feature abrupt change coefficient using multiple radar local features; determining a first local vibration energy gradient using multiple vibration local features; determining a first cooperative difference using the first radar local features and the first vibration local features; and determining a first local void anomaly degree using the first cooperative difference, the first local vibration energy gradient, and the first local radar feature abrupt change coefficient, thereby determining the interlayer void situation. This invention can effectively improve detection accuracy, achieve precise location of void areas, and enhance the accuracy of cooperative analysis.

Description

Technical Field

[0001] This invention relates to the field of testing technology, specifically to a device and method for detecting interlayer voids in cement concrete pavement. Background Technology

[0002] As a crucial transportation infrastructure, the structural integrity of cement concrete pavement directly impacts driving safety and service life. Over long-term use, interlayer delamination can easily occur between the pavement's structural layers, meaning gaps and separation form between the surface layer and the base layer, or between the base layer and the subbase layer. This delamination disrupts the stress transmission path, causing localized stress concentration, which in turn leads to a series of problems such as pavement cracking and reduced load-bearing capacity, seriously threatening driving safety.

[0003] Traditional inter-layer void detection technologies primarily rely on single detection methods, such as ground-penetrating radar or vibration testing. While these methods can initially identify the presence of void risks in road sections, they have significant limitations in practical applications: firstly, single detection methods are easily affected by environmental interference, leading to misjudgments; secondly, existing technologies struggle to accurately distinguish between overall void risk and specific void locations, failing to achieve precise localization from a global perspective. Particularly in small-scale void detection, existing technologies lack the ability to finely analyze the differences in characteristics between adjacent areas, making it difficult to accurately identify void boundaries.

[0004] Furthermore, existing detection methods often overlook the synergistic relationship between radar signals and vibration signals, leading to insufficient reliability of detection results. In practical engineering, interlayer voids simultaneously affect radar reflection characteristics and road vibration characteristics, but current technologies fail to effectively utilize this correlation between multiple physical quantities, resulting in low detection accuracy. At the same time, the criteria for determining void areas lack scientific basis, relying mainly on empirical thresholds, which are difficult to adapt to the specific conditions of different road sections.

[0005] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0006] In order to solve the existing technical problems, the purpose of this invention is to provide a device and method for detecting interlayer voids in cement concrete pavement, which has the advantages of improving detection accuracy, accurately locating void areas, and enhancing the collaborative analysis of multiple physical quantities.

[0007] To solve the above-mentioned technical problems, the technical solution adopted in this application is as follows: A method for detecting interlayer voids in cement concrete pavement is provided, comprising: acquiring first radar local features and first vibration local features of a first region of the cement concrete pavement; acquiring second radar local features and second vibration local features of a second region; and acquiring third radar local features and third vibration local features of a third region, wherein the cement concrete pavement is divided into at least three regions, and the first region is adjacent to the second region and the third region respectively; determining a local radar feature abrupt change coefficient using the first radar local features, the second radar local features, and the third radar local features; determining a local vibration energy gradient using the first vibration local features, the second vibration local features, and the third vibration local features; determining a first synergy difference of the first region using the first radar local features and the first vibration local features; and determining the local void anomaly degree of the first region using the first synergy difference, the local vibration energy gradient, and the local radar feature abrupt change coefficient, thereby determining the interlayer void situation of the cement concrete pavement.

[0008] In one embodiment of the present invention, obtaining the first radar local feature and the first vibration local feature of the first region of the cement concrete pavement includes: obtaining the first bimodal distance of the radar frame of the first region and the first quantization range of the radar reflection signal, and using the first bimodal distance and the first quantization range to determine the first bimodal spacing normalized value; obtaining the first bimodal height difference mean and the first maximum peak height mean of the radar frame of the first region, and using the first bimodal height difference mean and the first maximum peak height mean to determine the first bimodal height normalized value; using the first bimodal spacing normalized value and the first bimodal height normalized value to determine the first radar local feature; obtaining the first frequency band energy sum of the low-frequency energy of the first region, and obtaining the second frequency band energy sum of the full-frequency energy of the first region; using the first frequency band energy sum and the second frequency band energy sum to determine the first vibration local feature.

[0009] In one embodiment of the present invention, determining the first radar local feature using the normalized value of the bimodal spacing and the normalized value of the bimodal height includes: determining a positive correlation index using the difference between a first preset value and the normalized value of the bimodal height; determining the first radar local feature using the positive correlation index and the normalized value of the bimodal spacing; determining the first vibration local feature using the sum of the energy of the first frequency band and the sum of the energy of the second frequency band includes: obtaining the total number of sampling points of the first vibration local feature in the first region; determining the first vibration local feature using the total number of sampling points, the sum of the energy of the first frequency band, and the sum of the energy of the second frequency band.

[0010] In one embodiment of the present invention, determining the first local radar feature abrupt change coefficient of a first region using the first radar local feature, the second radar local feature, and the third radar local feature includes: obtaining a first change amplitude between the first radar local feature and the second radar local feature, and obtaining a second change amplitude between the first radar local feature and the third radar local feature; and determining the first local radar feature abrupt change coefficient using the first change amplitude and the second change amplitude.

[0011] In one embodiment of the present invention, determining the first local radar feature abrupt change coefficient using the first change amplitude and the second change amplitude includes: acquiring the global voiding anomaly features of the cement concrete pavement and determining the global normal fluctuation threshold of the adjacent bimodal differences of the radar frame; and determining the first local radar feature abrupt change coefficient using the global normal fluctuation threshold and the maximum value of the first change amplitude and the second change amplitude.

[0012] In one embodiment of the present invention, determining the first local vibration energy gradient of a first region using the first vibration local feature, the second vibration local feature, and the third vibration local feature includes: determining a first difference using the first vibration local feature and the second vibration local feature; determining a second difference using the first vibration local feature and the third vibration local feature; determining a directional consistency coefficient using the first difference and the second difference; and determining the first local vibration energy gradient using the directional consistency coefficient, the second difference, and a first length of the first region in response to the first vibration local feature being greater than the second vibration local feature.

[0013] In one embodiment of the present invention, determining the first coordination difference of the first region using the first radar local features and the first vibration local features includes: determining the first coordination coefficient of the first region using the first radar local features and the first vibration local features; determining the coordination average value of all regions of the cement concrete pavement using the first coordination coefficient; determining the third difference using the first coordination coefficient and the coordination average value; and determining the first coordination difference using the third difference and the coordination average value.

[0014] In one embodiment of the present invention, determining the first coordination coefficient of the first region using the first radar local features and the first vibration local features includes: obtaining the average value of radar local features and the average value of vibration local features corresponding to all regions of the cement concrete pavement; determining a first relative anomaly amplitude using the average value of radar local features and the first radar local features; determining a second relative anomaly amplitude using the average value of vibration local features and the first vibration local features; determining a first matching degree using the first relative anomaly amplitude, the second relative anomaly amplitude, and a second preset value; and determining the first coordination coefficient of the first region using the first matching degree, the first relative anomaly amplitude, and the second relative anomaly amplitude.

[0015] In one embodiment of the present invention, determining the first local voiding anomaly degree of the first region using the first synergy difference, the local vibration energy gradient, and the local radar feature mutation coefficient includes: acquiring global radar voiding characteristics and global vibration stiffness voiding characteristics of the cement concrete pavement; determining global voiding anomaly characteristics using the global radar voiding characteristics and the global vibration stiffness voiding characteristics, and determining global weights using the global voiding anomaly characteristics; determining the magnitude and component sum of the voiding feature vector using the first synergy difference, the first local vibration energy gradient, and the first local radar feature mutation coefficient; and determining the first local voiding anomaly degree of the first region using the magnitude, the component sum, and the global weights.

[0016] To address the aforementioned technical problems, another technical solution adopted in this application is to provide a device for detecting interlayer voids in cement concrete pavement, comprising: an acquisition module for acquiring first radar local features and first vibration local features of a first region of the cement concrete pavement, and acquiring second radar local features and second vibration local features of a second region, and third radar local features and third vibration local features of a third region, wherein the cement concrete pavement is divided into at least three regions, and the first region is adjacent to the second region and the third region respectively; a first determination module for determining a first local radar feature abrupt change coefficient using the first radar local features, the second radar local features, and the third radar local features; a second determination module for determining a first local vibration energy gradient using the first vibration local features, the second vibration local features, and the third vibration local features; a third determination module for determining a first synergy difference of the first region using the first radar local features and the first vibration local features; and a fourth determination module for determining a first local void anomaly degree of the first region using the first synergy difference, the first local vibration energy gradient, and the first local radar feature abrupt change coefficient, thereby determining the interlayer void situation of the cement concrete pavement.

[0017] The beneficial effects of the present invention are as follows: The method for detecting interlayer voids in cement concrete pavement provides a method that, by integrating radar and vibration characteristics into a multi-dimensional analysis, and combining the detection of abrupt changes in features of adjacent areas with the calculation of synergistic differences, achieves accurate identification of interlayer void areas, which has the advantages of improving detection accuracy and enhancing the synergistic analysis of multiple physical quantities. Attached Figure Description

[0018] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention 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 the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the interlayer void detection method for cement concrete pavement provided by the present invention.

[0020] Figure 2 This is a schematic diagram of the interlayer void detection device for cement concrete pavement provided by the present invention. Detailed Implementation

[0021] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a device and method for detecting interlayer voids in cement concrete pavement according to the present invention. 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.

[0022] 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 invention pertains.

[0023] In existing technologies, the detection of voids between layers in cement concrete pavements mainly relies on single detection methods or global data analysis, making it difficult to accurately locate local void areas. Traditional methods acquire overall data using ground-penetrating radar or vibration sensors and analyze abnormal indicators across the entire road section, but they cannot effectively distinguish the subtle differences between local voids and normal areas. For example, while global radar reflection signal analysis can identify bimodal characteristics, it cannot capture abrupt changes between adjacent areas; while vibration energy statistics can reflect stiffness attenuation, it is difficult to quantify local gradient changes in energy propagation. These limitations lead to the risk of misjudgment in the detection results, potentially misdiagnosing local anomalies as global problems or ignoring truly existing void areas, increasing maintenance costs and safety hazards.

[0024] To address the aforementioned issues, a detection method capable of balancing global risk assessment with local anomaly localization is needed. Considering the significant physical differences between vacant areas and adjacent normal areas, precise localization can be achieved by focusing on abrupt feature changes and coordinated anomalies between adjacent small areas based on global analysis. For example, the radar reflection signal of a vacant area exhibits a sudden jump at the boundary of adjacent areas, and vibration energy shows a steep attenuation along the propagation path; both types of features show synchronous anomalies at the boundary. Based on this, by dividing the road segment into multiple adjacent areas, extracting local radar and vibration features separately, and calculating the feature differences and coordination between adjacent areas, a multi-dimensional anomaly index can be constructed, thereby effectively distinguishing true vacant areas.

[0025] This application proposes a device and method for detecting interlayer voids in cement concrete pavement, which has the advantages of improving detection accuracy, accurately locating void areas, and enhancing the collaborative analysis of multiple physical quantities, thus effectively improving detection accuracy.

[0026] The specific scheme of the interlayer void detection method for cement concrete pavement provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0027] Please see Figure 1 The diagram illustrates a flowchart of the method for detecting interlayer voids in cement concrete pavement provided by the present invention.

[0028] like Figure 1 As shown, the method for detecting interlayer voids in cement concrete pavement includes the following steps:

[0029] S10. Obtain the first radar local feature and the first vibration local feature of the first region of the cement concrete pavement, and obtain the second radar local feature and the second vibration local feature of the second region, and the third radar local feature and the third vibration local feature of the third region, wherein the cement concrete pavement is divided into at least three regions, and the first region is adjacent to the second region and the third region respectively.

[0030] Among them, radar local features refer to quantitative indicators reflecting the bimodal shape of radar reflection signals within a region. Specifically, they can be calculated by combining the normalized values ​​of the bimodal spacing and the normalized values ​​of the bimodal height. That is, the first radar local feature is used to characterize the difference in interface reflection intensity caused by delamination in the first region, the second radar local feature is used to characterize the difference in interface reflection intensity caused by delamination in the second region, and the third radar local feature is used to characterize the difference in interface reflection intensity caused by delamination in the third region. Vibration local features refer to quantitative indicators reflecting the distribution of vibration energy within a region. Specifically, they can be calculated by the ratio of low-frequency energy to full-band energy. That is, the first vibration local feature is used to characterize the degree of stiffness attenuation caused by delamination in the first region, the second vibration local feature is used to characterize the degree of stiffness attenuation caused by delamination in the second region, and the third vibration local feature is used to characterize the degree of stiffness attenuation caused by delamination in the third region.

[0031] Specifically, the cement concrete pavement is divided into at least three regions, namely, at least a first region, a second region, and a third region, with the first region being adjacent to the second region and the third region, respectively. Then, for the first region, the second region, and the third region, quantitative indicators of the bimodal shape of the radar reflection signal within each region are obtained, namely, the first radar local feature, the second radar local feature, and the third radar local feature are obtained, respectively. Furthermore, for the first region, the second region, and the third region, quantitative indicators reflecting the vibration energy distribution within each region are obtained, namely, the first vibration local feature, the second vibration local feature, and the third vibration local feature are obtained, respectively.

[0032] S20. Using the local features of the first radar, the local features of the second radar, and the local features of the third radar, determine the first local radar feature mutation coefficient of the first region.

[0033] The first local radar feature mutation coefficient refers to the mutation amplitude of the radar feature difference between adjacent regions in the first region. Specifically, it can be calculated by comparing the change amplitude of radar features in adjacent regions with the global normal fluctuation threshold, and is used to identify local anomalies in interface reflection signals.

[0034] Specifically, since the first region is adjacent to the second and third regions respectively, that is, the first region is located in the middle of the second and third regions, the first local radar feature mutation coefficient of the middle first region can be determined based on the radar local features of each region; that is, the change range between the first radar local features and the second and third radar local features respectively is obtained, and then the first local radar feature mutation coefficient is determined.

[0035] S30. Using the first vibration local features, the second vibration local features and the third vibration local features, determine the first local vibration energy gradient of the first region.

[0036] The first local vibration energy gradient refers to the rate attenuation of vibration energy in the first region between adjacent regions. Specifically, it can be calculated by combining the directional consistency coefficient of the vibration characteristic difference between adjacent regions with the unit length difference rate, and is used to quantify the steep changes in the energy propagation path.

[0037] Specifically, the degree of change in vibration energy can be determined based on the differences between the local vibration characteristics of each region, that is, the degree of change between the first local vibration characteristic and the second and third local vibration characteristics can be obtained, and then the first local vibration energy gradient of the first region can be determined.

[0038] S40. Using the first radar local features and the first vibration local features, determine the first synergy difference in the first region.

[0039] The first synergy difference refers to the degree of synergy anomaly between the two types of features in the target region relative to the global benchmark. It can be a numerical value, specifically calculated by the relative difference between the regional synergy coefficient and the global synergy mean, and is used to verify the reliability of multi-feature synchronization anomalies.

[0040] Specifically, the coordination coefficient is calculated by using the first radar local features and the first vibration local features of the first region relative to the global mean, and then compared with the global coordination mean to obtain the first coordination difference.

[0041] S50. Using the first synergy difference, the first local vibration energy gradient, and the first local radar characteristic mutation coefficient, determine the first local void anomaly degree of the first region, and then determine the interlayer void situation of the cement concrete pavement.

[0042] Specifically, by integrating the local radar feature mutation coefficient, local vibration energy gradient, and first synergy difference, a void feature vector is constructed, the first local void anomaly degree of the first region is calculated, and then the interlayer void situation of the cement concrete pavement is determined.

[0043] Compared to existing technologies, traditional methods rely solely on single features or global statistical indicators, making it difficult to distinguish between local anomalies and noise interference. For example, analyzing radar bimodal characteristics alone might misjudge material inhomogeneity as voids, and statistically analyzing vibration energy alone might ignore the influence of the propagation path. This approach effectively eliminates isolated anomaly interference by introducing feature mutation analysis between adjacent regions and multi-feature collaborative verification. For instance, the local radar feature mutation coefficient is combined with a global threshold to filter random fluctuations, the local vibration energy gradient eliminates reverse differences through directional consistency, and collaborative differences verify the synchronicity of multi-dimensional anomalies. The combination of these three methods significantly improves detection accuracy.

[0044] Through the above technical solution, this application can accurately identify the specific location of voids between layers of cement concrete pavement, avoiding misjudging local problems as global risks. This method effectively distinguishes between actual void areas and interference factors such as material inhomogeneity through characteristic difference analysis of adjacent areas; reduces the probability of false alarms from a single sensor through multi-feature collaborative verification; and achieves risk classification and precise location through dynamic weight integration of global and local indicators, providing a reliable basis for maintenance operations.

[0045] In some embodiments, data acquisition may include the following steps: For interlayer detection of cement concrete pavement, the radar in the acquisition device is selected with a center frequency of 1.5-2.4 GHz. This frequency range can penetrate the pavement panel to reach the interlayer interface and can also clearly distinguish the reflection difference between normal bonding and voided air layers; the vibration detection device is selected to provide controllable impact load and synchronously acquire pavement vibration response. It needs to include a stable impact force application module (such as a drop hammer or pulse structure that can output a controllable impact force of 50-100kN adapted to the pavement stiffness) and a response acquisition module (equipped with an acceleration or displacement sensor, sampling frequency ≥200Hz, to capture low-frequency vibration energy changes caused by voiding).

[0046] To ensure that radar data and vibration data correspond in spatial location, the original continuous radar frame images are divided into independent detection units according to vibration point intervals (e.g., 10 meters / measuring point). A radar reflection signal curve is generated for each unit, and the distance between the two peaks and the height of the two peaks are extracted by peak-valley detection. The original vibration time-domain signal is converted into a frequency-domain signal using FFT (Fast Fourier Transform), and the energy ratio of the stiffness-sensitive low-frequency band to the total frequency band is statistically analyzed to obtain the proportion of vibration low-frequency energy for each detection unit.

[0047] In some embodiments, obtaining a first radar local feature and a first vibration local feature of a first region of a cement concrete pavement may include the following steps.

[0048] The first bimodal distance and the first quantization range of the radar reflection signal of the radar frame in the first region are obtained, and the normalized value of the first bimodal spacing is determined using the first bimodal distance and the first quantization range. The average height and maximum peak height of the first bimodal in the radar frame of the first region are obtained, and the normalized value of the first bimodal height is determined using the average height and maximum peak height. The normalized value of the first bimodal spacing and the normalized value of the first bimodal height are used to determine the local features of the first radar.

[0049] The first frequency band energy sum of the low-frequency energy of the first region is obtained, and the second frequency band energy sum of the full frequency band energy of the first region is obtained; the first vibration local characteristics are determined by using the first frequency band energy sum and the second frequency band energy sum.

[0050] The first double-peak spacing normalized value refers to the value obtained by standardizing the distance between the two reflection peaks in the radar reflection signal. Specifically, it can be achieved by dividing the first double-peak distance by the first quantization range of the radar reflection signal. This eliminates absolute numerical differences caused by different radar equipment or signal strengths, enhancing the comparability between different detection units. The first double-peak height normalized value is the ratio of the average height of the two reflection peaks to the maximum peak height after standardization. Specifically, it can be achieved by dividing the average height of the first double-peak by the maximum peak height. This quantifies the relative degree of the double-peak height difference, avoiding misjudgments caused by fluctuations in absolute signal strength. The first radar local feature refers to the normalized index of the comprehensive double-peak spacing and height difference in the first region. Specifically, it can be calculated by combining the first double-peak spacing normalized value and the first double-peak height normalized value. This characterizes the intensity of the double-interface reflection anomaly caused by voiding in the radar signal. The first frequency band energy summation refers to the cumulative energy value of the vibration signal in the low-frequency band. Specifically, it can be obtained by integrating the vibration signal in the 20-50Hz frequency band after performing a fast Fourier transform. This reflects the low-frequency energy accumulation phenomenon caused by the decrease in road surface stiffness due to voiding. The total energy of the second frequency band refers to the cumulative energy of the vibration signal across the entire frequency band. Specifically, it can be obtained by integrating the energy of the vibration signal in the 0-500Hz frequency band. This energy is used to calculate the proportion of low-frequency energy, thereby eliminating the difference in the absolute value of vibration energy between different detection units.

[0051] Specifically, when acquiring the first radar local features, the first double-peak distance and the first quantization range corresponding to the first region are first extracted from the radar frame, i.e., from the radar reflection signal. The influence of equipment differences on the double-peak distance is eliminated through normalization processing to obtain the normalized value of the first double-peak distance. At the same time, the average value of the first double-peak height difference and the average value of the first maximum peak height are extracted. The relative degree of the double-peak height difference is quantified through normalization processing to obtain the normalized value of the first double-peak height. Finally, the two types of normalization indicators are combined to generate the first radar local features.

[0052] When acquiring the first local vibration feature, the total energy of the first frequency band of low-frequency energy in the first region is first acquired, and the total energy of the second frequency band of full-frequency energy in the first region is also acquired. The ratio of low-frequency energy to full-frequency energy is calculated, and this ratio is then used to determine the first local vibration feature, eliminating the influence of absolute energy fluctuations in the vibration signal and accurately reflecting the change in the proportion of low-frequency energy caused by delamination. Through these steps, the raw radar and vibration signal data can be transformed into standardized features that can be compared laterally, providing a consistent data foundation for subsequent local delamination detection.

[0053] Through the above technical solution, this application can solve the problem of incomparable feature data caused by equipment differences or environmental interference. By normalization processing and energy ratio calculation, the robustness of radar and vibration features is enhanced, making the local void detection results more stable and reliable. At the same time, by fusing radar features with differences in the spacing and height of the double peaks, as well as vibration features with low-frequency energy ratio, multi-dimensional signal anomalies caused by voiding can be captured more comprehensively, avoiding the limitations of single feature detection.

[0054] Furthermore, determining the local features of the first radar using the normalized values ​​of the first bimodal spacing and the first bimodal height may include the following steps.

[0055] The difference between the first preset value and the first bimodal height normalized value is used to determine the positive correlation index; the positive correlation index and the first bimodal spacing normalized value are used to determine the local characteristics of the first radar.

[0056] Among them, the positive correlation index refers to the value obtained by subtracting the normalized value of the first bimodal height from the preset value. Specifically, it can be achieved by setting the preset value to 1 and performing the subtraction operation. This index is used to reflect the degree of deviation of the bimodal height from the normal state. The larger the value, the smaller the difference in bimodal height and the higher the possibility of detachment.

[0057] Specifically, when determining the first local radar feature, the normalized value of the first bimodal height is first converted into a positive correlation index. This index is then multiplied by the normalized value of the first bimodal spacing to obtain the local radar feature. This calculation method can simultaneously consider the influence of the bimodal spacing and height difference on the clearance detection, avoiding misjudgments caused by noise interference from a single index.

[0058] In some embodiments, the global radar clearance characteristics of the cement concrete pavement can be obtained first. Specifically, this involves: obtaining the global mean distance between the two peaks of all radar frames of the cement concrete pavement and the global quantization range of the radar reflection signal; then using the global mean distance between the two peaks and the global quantization range to determine the normalized value of the global two-peak spacing; obtaining the global mean height difference between the two peaks and the global mean maximum peak height of the cement concrete pavement; then using the global mean height difference between the two peaks and the global mean maximum peak height to determine the normalized value of the global two-peak height. Finally, using the normalized value of the global two-peak spacing and the normalized value of the global two-peak height, the global radar clearance characteristics are determined.

[0059] In this process, voids create double-reflection interfaces between the lower surface of the concrete and the air layer, and between the air layer and the upper surface of the base layer. Radar waves reflected through these interfaces produce two reflected signals with a propagation time difference. After receiving the reflected signals from the sampling points, the acquisition equipment can determine the void situation by comparing signal characteristics: in dense areas without voids, the radar wave propagation path is relatively stable; while in areas with voids, the reflection signal curve exhibits a distinct bimodal characteristic. When analyzing the curve, the initial reflected signal directed towards the ground surface should be ignored, and the first two largest peaks should be extracted. The radar wave propagation time difference corresponding to these two peaks is essentially the round-trip time difference of the reflected waves between the two interfaces. The larger this time difference, the thicker the air layer and the greater the void; the wider the void area, the larger the exposed area of ​​the base layer interface, the stronger the reflected signal, and the closer the heights of the two peaks. Therefore, the smaller the height difference, the wider the void area.

[0060] Specifically, in order to eliminate the absolute numerical differences in radar signals across different road sections, two types of indicators are normalized: the global bimodal distance normalized value. The mean global bimodal distance of all radar frames across the entire road segment ( The ratio of the radar reflection signal to the global quantization range of the radar reflection signal (e.g., 256, depending on the radar model): Global bimodal height normalized value The mean of the global bimodal height difference for the entire road segment ( ) and the global maximum peak average ( The ratio of ) The significance of the clearance feature must simultaneously satisfy the conditions of large spacing and small height difference. Therefore, the product of the two types of normalized indices is the global radar clearance feature H.

[0061]

[0062] Since the smaller the height difference, the more significant the gap is, the formula uses... The inverse correlation is converted into a positive correlation index, with the first preset value being 1.

[0063] It is understandable that the normalized value of the first bimodal distance corresponding to the first region is determined by the average first bimodal distance and the first quantization range, the normalized value of the first bimodal height is determined by the average first bimodal height difference and the average first maximum peak height, and the first radar local feature is determined by the normalized value of the first bimodal distance and the normalized value of the first bimodal height. The calculation process is the same as the calculation of determining the global radar clearance feature using the normalized values ​​of the global bimodal distance and the global bimodal height, and will not be elaborated here.

[0064] Furthermore, the normalized values ​​of the second bimodal distance, the second bimodal height, and the second radar local features corresponding to the second region, and the normalized values ​​of the third bimodal distance, the third bimodal height, and the third radar local features corresponding to the third region can also be calculated using the global radar clearance features, which will not be elaborated here.

[0065] Furthermore, determining the local characteristics of the first vibration by utilizing the sum of energy in the first frequency band and the sum of energy in the second frequency band may include the following steps.

[0066] Obtain the total number of sampling points for the first vibration local feature in the first region; determine the first vibration local feature using the total number of sampling points, the sum of energy in the first frequency band, and the sum of energy in the second frequency band.

[0067] The total number of sampling points refers to the number of vibration signal samples in the time dimension, which can be achieved by counting the total number of points in the time domain signal. This parameter is used to eliminate the influence of different sampling durations on the total energy and ensure the stability of the calculation of local vibration features. The total energy of the first frequency band and the total energy of the second frequency band refer to the cumulative vibration energy values ​​of the low-frequency band and the full frequency band, respectively.

[0068] Specifically, when determining the first local vibration feature, the ratio of the sum of low-frequency energy to the sum of energy across the entire frequency band is multiplied by the total number of sampling points. This ensures that the calculation results are unaffected by fluctuations in sampling duration, thus more stably characterizing the distribution of vibration energy. Through these steps, key features reflecting interlayer voiding can be effectively extracted, providing reliable input for subsequent calculations of local anomalies.

[0069] For example, the force on the surface layer could originally be transferred to the base layer, but the air layer created after the surface layer is de-voiced blocks this force transfer, forcing the surface layer to bear the load alone, resulting in a decrease in the overall stiffness of the road surface. Road surface stiffness is positively correlated with the natural frequency of vibration (lower stiffness, lower natural frequency), and changes in the natural frequency alter the distribution of vibration energy: under normal road surface stiffness, vibration energy is concentrated in the mid-to-high frequency range; after de-voicing, the reduced stiffness lowers the natural frequency, causing more vibration energy to concentrate in the low-frequency range. For a single sampling point i, the total energy of the first frequency band of the low-frequency energy (e.g., 20-50Hz) in the first region is... The sum of the energy of the second frequency band from the full-band energy of the first region (e.g., 0-500Hz) is: Then the low-frequency proportion of this sampling point is The global vibration stiffness decoupling characteristic V needs to reflect the overall stiffness attenuation of the entire road section; therefore, the average of the low-frequency proportions of all sampling points (N in total) is taken:

[0070] .

[0071] Where N is the total number of sampling points.

[0072] The determination of global de-energization must simultaneously ensure the abnormality levels of the two signals mentioned above, with H... If V exhibits two types of anomaly intensity, then the global void anomalous feature G is:

[0073]

[0074] Wherein, H represents the global radar clearance characteristic; This indicates the global vibration stiffness decoupling characteristic.

[0075] As a global feature of de-cashing, the numerical value can intuitively reflect the global de-cashing risk: the higher the value, the greater the intensity of the dual-feature anomaly, which can help verify the credibility of local anomalies and make local positioning fit the actual risk of the entire road section.

[0076] Through the above technical solution, this application can more accurately quantify the abnormal features in radar reflection signals and vibration signals, effectively distinguish between real void areas and local noise interference, and improve the accuracy of interlayer void detection. At the same time, this solution integrates multi-dimensional features through mathematical operations, reducing data processing complexity and providing an efficient computational foundation for large-scale pavement detection.

[0077] In some embodiments, determining the first local radar feature abrupt change coefficient of a first region using a first radar local feature, a second radar local feature, and a third radar local feature may include the following steps.

[0078] Obtain the first variation amplitude between the first radar local feature and the second radar local feature, and obtain the second variation amplitude between the first radar local feature and the third radar local feature; use the first variation amplitude and the second variation amplitude to determine the abrupt change coefficient of the first local radar feature.

[0079] The first variation amplitude refers to the degree of difference between the first radar local feature and the adjacent second radar local feature, which can be represented by the absolute value of their difference, reflecting the intensity of abrupt changes in the radar reflection signal between adjacent regions. The second variation amplitude refers to the degree of difference between the first radar local feature and the adjacent third radar local feature, which can also be represented by the absolute value of their difference, reflecting the intensity of abrupt changes in the radar reflection signal in another adjacent direction. The first local radar feature abrupt change coefficient is a quantification index of abrupt changes obtained by combining the differences in radar features in two adjacent directions. It can be represented by comparing the maximum value of the two variation amplitudes with a preset threshold, used to identify abnormal abrupt changes in the radar reflection signal caused by gaps in the radar signature.

[0080] Specifically, during the detection process, for the first region to be judged, the differences in radar local features between it and the adjacent second and third regions are calculated. By extracting the amplitude of change in two directions—that is, obtaining the first amplitude and the second amplitude of change respectively—the abrupt changes in radar reflection signals between adjacent regions can be comprehensively captured. When the amplitude of change in a certain direction significantly exceeds the normal fluctuation range, it indicates that there may be an interface break caused by a gap in that direction. By combining the maximum amplitude of change in two directions, the limitations of single-direction detection can be effectively avoided, and the robustness of abrupt change identification can be improved. For example, when the first amplitude of change is 0.8 and the second amplitude of change is 0.5, the maximum value of 0.8 is preferentially selected for calculating the abrupt change coefficient, thereby focusing on the most significant abnormal signal.

[0081] Through the above technical solution, this application can more accurately identify the sudden change phenomenon of radar reflection signal caused by interlayer delamination, effectively distinguish the real delamination area from local data fluctuations through multi-directional difference comparison, provide reliable radar feature change indicators for subsequent collaborative judgment, thereby improving the accuracy of delamination area positioning and the credibility of detection results.

[0082] Furthermore, determining the first local radar feature abrupt change coefficient using the first and second change amplitudes may include the following steps.

[0083] The global voiding anomaly features of cement concrete pavement are obtained, and the global normal fluctuation threshold of the difference between adjacent bimodal peaks in radar frames is determined. The first local radar feature mutation coefficient is determined by using the global normal fluctuation threshold and the maximum value of the first and second variation amplitudes.

[0084] Among them, the global de-airing anomaly characteristic refers to a comprehensive indicator reflecting the overall de-airing risk of the entire road segment. Specifically, it can be calculated by multiplying the normalized value of the distance between the radar peaks and the normalized value of the height of the peaks across the entire road segment, combined with the average proportion of low-frequency vibration energy, and is used to measure the intensity of the global de-airing anomaly. The normal fluctuation threshold refers to the reasonable fluctuation range of radar characteristic differences between adjacent areas. Specifically, it can be calculated by statistically analyzing the standard deviation of radar characteristic differences between all adjacent areas across the entire road segment, for example, taking twice the standard deviation as the threshold, to distinguish between normal fluctuations and abnormal abrupt changes. The maximum value between the first and second variation amplitudes refers to the larger of the absolute values ​​of radar characteristic differences between the area to be judged and its preceding and following adjacent areas. Specifically, it can be calculated by taking the maximum value of the absolute values ​​of the local radar characteristic differences between the area to be judged and its adjacent areas, and is used to capture the most significant local abrupt changes.

[0085] Specifically, in determining the local radar feature mutation coefficient, firstly, based on the global voiding anomaly characteristics of the cement concrete pavement, the distribution of radar feature differences between adjacent areas across the entire road segment is statistically analyzed, i.e., the global normal fluctuation threshold of adjacent bimodal differences in radar frames is obtained; for example, the standard deviation of differences between all adjacent areas is calculated, and twice the standard deviation is set as the normal fluctuation threshold. Subsequently, the absolute values ​​of radar feature differences between the area to be judged and its preceding and following adjacent areas are calculated, and the larger absolute value is selected as the maximum variation amplitude, i.e., the maximum value between the first and second variation amplitudes is taken as the maximum variation amplitude. Finally, the ratio of the maximum variation amplitude to the global normal fluctuation threshold is calculated to obtain the first local radar feature mutation coefficient. When this coefficient is greater than 1, it indicates that the radar feature differences between the area to be judged and its adjacent areas exceed the normal fluctuation range, indicating a significant mutation.

[0086] For example, to detect changes between adjacent road surface areas, the road surface to be detected is divided into several independent detection units (e.g., 1-2 meters / unit), that is, divided into several regions, ensuring that each unit can synchronously match a set of first radar local features. Local characteristics of the first vibration .

[0087] The change from continuity to fracture at the interlayer interface caused by voiding is reflected in the difference in interface characteristics between the unit to be judged (first region) and its directly adjacent units before and after it (second and third regions) (a total of 3 units constituting a fixed local area). In normal pavement, the interlayer interface is continuous, and the difference in radar bimodal characteristics between the unit to be judged (first region) and its directly adjacent units before and after it (second and third regions) is gradual. If voiding exists in the unit to be judged (first region), the difference in radar bimodal characteristics with at least one adjacent unit (second or third region) will be significant. For the radar bimodal characteristic (H), the absolute value of the difference between the unit to be judged (first region) and its directly adjacent units before and after it (adjacent units j-1 and j+1, i.e., second and third regions) is calculated to reflect the magnitude of the interface characteristic change: First magnitude of change: Second variation range Based on the global gap-out anomaly feature G statistics, the global normal fluctuation threshold of the adjacent differences in radar bimodal features is obtained. (e.g., 2 standard deviations) represents the maximum reasonable range of characteristic differences between adjacent units in a normal road surface; then the final local radar characteristic mutation coefficient of the first region is... for:

[0088]

[0089] in Because if the unit to be judged has a gap, at least one adjacent difference will exceed the maximum reasonable range. Therefore, the larger of the two is selected to reflect the most significant single adjacent difference, and then divided by the global normal fluctuation threshold. This transforms absolute differences into comparative values ​​that represent relatively normal fluctuations.

[0090] It should be noted that, to ensure the calculation results are meaningful, in this embodiment of the application, when performing fractional operations, if the denominator is 0, a parameter adjustment factor can be added to the denominator to prevent the denominator from being 0. This parameter adjustment factor is a very small positive number. For example, the value of this parameter adjustment factor can be 0.01. Its specific value can be set by the implementer according to the actual situation, and this embodiment of the application does not impose a specific limitation.

[0091] Through the above technical solution, this application can accurately identify radar feature abrupt change regions caused by interlayer gaps. By combining a global fluctuation threshold with the maximum adjacent difference, it reduces misjudgments caused by local noise or normal fluctuations, thereby improving the reliability of gap region positioning. Simultaneously, this method quantifies the ratio of difference to a threshold, intuitively reflecting the degree of abrupt change and providing a basis for subsequent gap degree classification.

[0092] In some embodiments, determining the first local vibration energy gradient of a first region using a first vibration local feature, a second vibration local feature, and a third vibration local feature may include the following steps.

[0093] The first difference is determined using the first and second local vibration features; the second difference is determined using the first and third local vibration features; and the directional consistency coefficient is determined using the first and second differences.

[0094] In response to the fact that the first local vibration feature is greater than the second local vibration feature, the first local vibration energy gradient is determined using the directional consistency coefficient, the second difference, and the first length of the first region.

[0095] The first difference refers to the difference between the first local vibration feature and the adjacent second local vibration feature. This can be achieved by subtracting the vibration feature values ​​of the two regions, reflecting the degree of abrupt change in vibration energy between adjacent regions. The second difference refers to the difference between the first local vibration feature and the adjacent third local vibration feature. This can also be achieved by subtracting the vibration feature values ​​of the two regions, reflecting the trend of vibration energy change in another adjacent direction. The directional consistency coefficient is a quantitative indicator of whether the directions of change of the two differences are consistent. This can be achieved by multiplying the results using a sign function, used to eliminate random fluctuation interference. The first length of the first region refers to the actual length of the road surface corresponding to that region. This can be achieved using the preset unit length parameter when dividing the detection unit, used to convert the vibration energy difference into a gradient value per unit distance.

[0096] Specifically, after the regions corresponding to the detection units are divided (one detection unit corresponds to one region), the difference in vibration characteristics between the unit to be judged and two adjacent units is calculated to determine whether the direction of energy change is consistent. When the first local vibration characteristic is greater than the second local vibration characteristic, it indicates that the vibration energy undergoes abnormal attenuation during its propagation to the third region. At this time, the second difference is weighted by combining the directional consistency coefficient to obtain the directional consistency coefficient, eliminating the interference caused by reverse fluctuations. Then, based on the unit length, the weighted difference is converted into an energy gradient value per unit distance, thus obtaining the first local vibration energy gradient. The first local vibration energy gradient can characterize the steep attenuation characteristics of vibration energy at the delamination boundary, thereby effectively distinguishing between normal energy dissipation and abnormal attenuation caused by delamination.

[0097] For example, in a normal road surface area without voids, vibration energy dissipates uniformly over the propagation distance, and the difference in the proportion of low-frequency vibration between adjacent units fluctuates randomly within a small range. In voided areas, the stiffness of the air layer is lower than that of the road surface material, leading to a larger difference in the proportion of low-frequency vibration between adjacent units. This change is not a random fluctuation, but rather a continuous increase in the proportion of low-frequency vibration between consecutive adjacent units along the boundary from the normal area to the voided area: the proportion of low-frequency vibration is small in the normal area due to weak energy absorption; the closer to or into the voided area, the stronger the energy absorption, and the larger the proportion of low-frequency vibration. Abnormalities in adjacent detection units... The difference rate of the proportion of low-frequency vibrations per unit length (the road surface length corresponding to each unit) between adjacent units was obtained; the directional consistency coefficient was also obtained. The calculation is as follows:

[0098] .

[0099] in, This is a sign function used to indicate the sign of the internal parameters; The first difference, The second difference, This is the first local characteristic of vibration. This is a local characteristic of the second vibration. The third vibration local characteristic, when and It is valid when the directions are consistent.

[0100] The first local vibration energy gradient for:

[0101]

[0102] in, This is the first length of the first region.

[0103] Through the above technical solution, this application can accurately capture the continuous attenuation characteristics of vibration energy along a specific direction at the boundary of the voided area, avoid misjudgment caused by local random fluctuations, and improve the reliability of interlayer void detection.

[0104] In some embodiments, determining a first synergy difference in a first region using a first radar local feature and a first vibration local feature may include the following steps.

[0105] Using the first radar local features and the first vibration local features, the first coordination coefficient of the first region is determined; using the first coordination coefficient, the coordination average value of all regions of the cement concrete pavement is determined; using the first coordination coefficient and the coordination average value, the third difference value is determined; using the third difference value and the coordination average value, the first coordination difference value is determined.

[0106] The first coordination coefficient refers to the degree of coordination anomaly between radar and vibration features in a local area. Specifically, it can be calculated by comparing the relative anomaly amplitudes of local radar features and the global average, and the relative anomaly amplitudes of local vibration features and the global average, combined with their matching degree. This coefficient quantifies the degree of synchronization anomaly between the two features caused by gaps in detection. The coordination average value refers to the mean of the coordination coefficients across all areas. Specifically, it can be calculated by taking the arithmetic mean of the coordination coefficients of all detection units, serving as a reference value for determining whether local coordination deviates from the global benchmark. The third difference refers to the difference between the local coordination coefficient and the global coordination average value, specifically achieved through subtraction, used to quantify the magnitude of the deviation of coordination from the benchmark in that area.

[0107] Specifically, the method first calculates the anomalous amplitudes of the first radar local feature and the first vibration local feature of the first region to be judged relative to the global average, and generates a first coordination coefficient by combining the matching degree of the two, reflecting the degree of synchronous anomaly of the two features in the region. Next, the mean coordination coefficient of all regions is calculated, i.e., the coordination average value, as a benchmark value. The difference between the first coordination coefficient of the first region to be judged and the coordination average value is obtained as a third difference value. The first coordination difference is then calculated by the ratio of this difference value to the coordination average value. For example, when the coordination coefficient of a region is significantly higher than the benchmark value, its coordination difference will show a positive increase, indicating that there is a coordination anomaly caused by gap separation in that region. Thus, this method effectively distinguishes between real gap separation regions and local fluctuations caused by random noise by establishing a comparison mechanism between the global benchmark and local differences.

[0108] Through the above technical solution, this application can effectively identify areas of coordinated abnormality in radar and vibration characteristics caused by interlayer gaps, and eliminate misjudgments caused by random fluctuations of a single feature. For example, when there is equipment signal interference, this method can accurately distinguish between real gaps and temporary interference by determining whether the two features deviate from the reference value synchronously, thereby improving the reliability of local gap location.

[0109] Furthermore, determining the first coordination coefficient of the first region using the first radar local features and the first vibration local features may include the following steps.

[0110] The average value of radar local features and the average value of vibration local features are obtained for all areas of the cement concrete pavement. The first relative anomaly amplitude is determined using the average value of radar local features and the first radar local features, and the second relative anomaly amplitude is determined using the average value of vibration local features and the first vibration local features. The first matching degree is determined using the first relative anomaly amplitude, the second relative anomaly amplitude, and the second preset value. The first coordination coefficient of the first region is determined using the first matching degree, the first relative anomaly amplitude, and the second relative anomaly amplitude.

[0111] The radar local feature average value refers to the average value of the normalized index of radar bi-peak spacing and height for all detection units across the entire road segment. Specifically, it can be achieved by summing the radar local features of all units and dividing by the total number of units, reflecting the normal baseline level of radar characteristics across the entire road segment. The vibration local feature average value refers to the average value of the proportion of low-frequency vibration energy for all detection units across the entire road segment. Specifically, it can be achieved by summing the vibration local features of all units and dividing by the total number of units, reflecting the normal baseline level of vibration characteristics across the entire road segment. The first relative anomaly amplitude refers to the degree of difference between the radar local features of the unit to be judged and the average value for the entire road segment. Specifically, it can be achieved by subtracting the average value for the entire road segment from the radar local features of the unit to be judged and taking the absolute value, quantifying the magnitude of the deviation of the unit's radar features from the normal baseline. The second relative anomaly amplitude refers to the degree of difference between the vibration local features of the unit to be judged and the average value for the entire road segment. Specifically, it can be achieved by subtracting the average value for the entire road segment from the vibration local features of the unit to be judged and taking the absolute value, quantifying the magnitude of the deviation of the unit's vibration features from the normal baseline. The first matching degree refers to the consistency index of the abnormal amplitude of radar and vibration characteristics. Specifically, it can be determined by the absolute value of the difference between two relative abnormal amplitudes, and is used to measure the degree of synchronization between the two types of characteristic anomalies. The first coordination coefficient refers to the strength index of the decoupling coordination due to the combination of abnormal amplitude and matching degree. Specifically, it can be achieved by taking the square root of the product of matching degree and two relative abnormal amplitudes, and is used to reflect the comprehensive strength of the dual-characteristic coordinated anomaly caused by decoupling in this unit.

[0112] Specifically, when determining the coordination coefficient, the anomalous amplitude of the radar and vibration local features of the unit to be judged relative to the average value of the entire road segment is first calculated. This involves obtaining the average radar local features and the average vibration local features corresponding to all areas of the cement concrete pavement. Then, using the average radar local features and the first radar local features of the first area, the first anomalous amplitude is determined, and using the average vibration local features and the first vibration local features of the first area, the second anomalous amplitude is determined. Directional differences are eliminated through absolute value calculations. Subsequently, the ratio of the two anomalous amplitudes is limited to a preset range to avoid a single feature dominating the matching degree calculation. The closer the matching degree is to 1, the closer the anomalous amplitudes of the two types of features are, and the stronger the coordination caused by the gap. Finally, the matching degree is multiplied by the geometric mean of the anomalous amplitudes to obtain the first coordination coefficient of the first area, which simultaneously considers the anomalous intensity and synchronicity. Thus, single-feature anomalies caused by random fluctuations in normal areas are suppressed due to low matching degrees, while the dual-feature coordinated anomalies in the actual gap areas are effectively amplified.

[0113] For example, when a gap occurs, the data collected by both devices will show the above-mentioned anomalies. Since a single feature anomaly might be due to noise interference, the first radar local feature should be considered. and the local characteristics of the first vibration It is only reliable when synchronization is abnormal. The average value of local radar features is obtained by using the median of the features of all cells as the normal baseline. and the average value of local vibration characteristics Based on this, only elements whose radar and vibration characteristics both exceed their respective normal baselines are considered as potential air-decoupling related elements.

[0114] For a potential unit, i.e., for the first region, calculate the magnitude of the first relevant anomaly:

[0115] .

[0116] Calculate the magnitude of the second relevant anomaly:

[0117] .

[0118] in, The magnitude of the first relevant anomaly. This represents the second relevant abnormal magnitude.

[0119] Because the impact of voiding on the road surface is synchronous, if a certain unit truly experiences an anomaly due to voiding, the abnormal amplitudes of radar and vibration should be roughly equivalent. Therefore, the first degree of matching between the two is... The calculation is as follows:

[0120] .

[0121] The second preset value is 1, and the first matching degree is... Greater than 1.

[0122] First matching degree The closer the value is to 1, the more consistent the magnitude of the feature anomalies. The cooperative strength of a single unit needs to reflect both the magnitude of the anomalies and the degree of matching; therefore, it is defined as follows: Represents the first synergy coefficient of unit j :

[0123] .

[0124] First Coordination Coefficient The larger the value, the stronger the de-vacancy of the unit.

[0125] Calculate the potential cooperative units of the road surface to be detected Average value, i.e. k is The number of units.

[0126] It is the quantization of the reflection signal at the interface between pavement layers, and its numerical variation is related to the continuity of the interface; This quantifies the energy loss due to vibration propagation along the road surface. The numerical change reflects the energy absorption intensity. Since their generation mechanisms differ, their trends in normal areas are independent and asynchronous. However, in areas with air gaps, the air layer formed by these gaps acts as both a radar-identifiable reflective interface and a vibration energy absorption zone, thus exhibiting a coordinated change. Therefore, the first difference in the coordination of the unit j to be judged can be quantified by calculating the percentage by which the coordination of a single unit exceeds the benchmark. That is, the first synergy difference in the first region. The calculation is as follows:

[0127]

[0128] in, This represents the first coordination coefficient of the unit j to be detected; This represents the collaborative average value of all units of the road surface to be tested, and serves as the baseline value. Used to find the maximum element of a vector or matrix.

[0129] Because the synergy of the delimited boundary will only be higher than the global average, and will not be lower than the global average, therefore... Only positive differences (above the baseline) are retained to avoid interference from normal cells.

[0130] Through the above technical solution, this application can accurately identify areas where radar and vibration characteristics are synchronously abnormal due to voiding, avoiding misjudgment caused by noise or local interference in single feature detection, and improving the reliability of interlayer voiding detection. Simultaneously, through matching degree constraints, the weights of the two types of features in different road sections can be adaptively adjusted to adapt to complex and ever-changing actual detection environments.

[0131] In some embodiments, the determination of a first local voiding anomaly degree in the first region is made using a first synergy difference, a first local vibration energy gradient, and a first local radar characteristic abrupt change coefficient, including:

[0132] The global radar voiding characteristics and global vibration stiffness voiding characteristics of cement concrete pavement are obtained; global voiding anomaly characteristics are determined using the global radar voiding characteristics and global vibration stiffness voiding characteristics, and global weights are determined using the global voiding anomaly characteristics; the magnitude and component sum of the voiding feature vector are determined using the first synergy difference, the first local vibration energy gradient, and the first local radar feature mutation coefficient; the first local voiding anomaly degree of the first region is determined using the magnitude, component sum, and global weights.

[0133] The global radar gap-out feature refers to the overall degree of gap-out anomaly calculated based on the normalized indices of the bi-peak spacing and height difference of the radar reflection signals across the entire road segment. Specifically, it can be achieved by calculating the ratio of the average bi-peak spacing to the quantization range and the ratio of the average bi-peak height difference to the average maximum peak height for all radar frames across the entire road segment, multiplying these ratios, and taking the square root. This feature reflects the abnormal intensity of radar reflection signals caused by air layers across the entire road segment. The global vibration stiffness gap-out feature refers to the overall stiffness attenuation calculated based on the average proportion of low-frequency vibration energy across the entire road segment. Specifically, it can be achieved by statistically analyzing the average ratio of low-frequency energy to total energy at all sampling points. This feature quantifies the abnormal distribution of vibration energy across the entire road segment caused by inter-layer gap-out. The global weight refers to the adjustment coefficient for local anomaly degree based on the global gap-out anomaly feature. Specifically, it can be achieved by exponentially calculating the ratio of the global gap-out anomaly feature to a preset benchmark value. This weight is used to enhance the credibility of local anomalies when the global risk is high and to suppress local misjudgments when the global risk is low. The detached feature vector is a three-dimensional vector composed of the first synergy difference, the local vibration energy gradient, and the local radar feature mutation coefficient. Specifically, it can be implemented by performing mathematical operations on these three components as vector elements. The modulus is used to measure the overall intensity of each feature anomaly, and the sum of the components characterizes the degree of synergy between features.

[0134] Specifically, when determining the local anomaly degree of de-cabling, the global de-cabling anomaly feature is first calculated by multiplying the global radar de-cabling feature and the global vibration stiffness de-cabling feature, thus characterizing the overall risk level of the entire road section. This feature is then converted into a global weight using an exponential function, allowing the global risk level to influence local assessments non-linearly. Next, the first synergy difference of the three local features, the first local vibration energy gradient, and the first local radar feature mutation coefficient are constructed into a three-dimensional vector, i.e., a de-cabling feature vector. Its magnitude and component sum are calculated for each feature. The magnitude reflects the comprehensive intensity of each anomaly, while the component sum reflects the synergy between features. Finally, the ratio of the magnitude to the component sum is multiplied by the global weight to obtain the first local de-cabling anomaly degree corresponding to the first region. For example, the magnitude can be calculated using the Euclidean distance formula, and the component sum can be calculated by directly adding the three features. The closer the ratio is to 1, the higher the feature synergy.

[0135] For example, the decoupling phenomenon exhibits anomalies in three dimensions: radar signal, vibration energy, and the synergy between the two. Therefore, these are integrated and defined as a decoupling feature vector. To avoid misjudgment caused by a single feature, we first... Identify the most significant anomalous signal; the essence of the void is multi-feature synchronous anomaly, which requires quantifying the synergy of the three, achieved through the ratio of the sum of the components of the feature vector to the magnitude.

[0136] The sum of the components is: .

[0137] Length of the module: .

[0138] in, A value greater than 0 indicates greater synergy; the closer the ratio is to 1, the more coordinated the collaboration.

[0139] Furthermore, local anomalies should be assessed within the context of overall risk. If there is no risk of derailment across the entire route, local false anomalies should be forcibly excluded. If overall risk exists, the weighting should increase slowly as the risk rises, and appropriate measures should be taken. As a global weight, the first local void anomaly degree The calculation is as follows:

[0140]

[0141] in, The first local voiding anomaly degree, This is the global weight.

[0142] Furthermore, the degree of the first local void anomaly was obtained. Then, a threshold T is set to determine if the road surface is clear (T can be based on the normal road surface). The 95th percentile (the criterion for determination) is as follows: The unit was determined to be free of voids. Preliminary assessment indicates that the unit has a void, and The higher the value, the more severe the devitrification (e.g., Mild desiccation, (This indicates severe voiding). Considering that voids are usually distributed in sheets, it is also necessary to perform linkage verification on adjacent elements. If three or more adjacent elements all meet the requirements... This confirms the authenticity of the voided area.

[0143] For units and regions identified as having vacancy, it is necessary to follow the procedures outlined above. Treatment plans are formulated based on the size of the affected area: units with minor voiding are included in the road surface periodic monitoring list, and the development of voiding is reviewed every 3-6 months; units with severe voiding are marked as key maintenance targets and reported to the maintenance department for targeted repair work.

[0144] Compared to existing technologies, current methods typically rely on a single feature or a simple linear combination to identify local anomalies, failing to consider the dynamic adjustment of global risk to local judgments or quantify the synergy between multiple features. For example, traditional methods may directly set a fixed threshold for the radar feature mutation coefficient; when the mutation coefficient exceeds the threshold, it is judged as a void, but without combining vibration energy gradient and synergy differences for comprehensive judgment. This scheme introduces global weights to achieve adaptive risk adjustment and uses vector operations to simultaneously evaluate feature strength and synergy, effectively avoiding misjudgments caused by local noise or single feature anomalies.

[0145] Through the above technical solution, this application can dynamically adjust the local judgment threshold according to the risk level of the entire road segment, improving the sensitivity to weak abnormal signals when the global risk is high and suppressing noise interference when the global risk is low. Simultaneously, by quantifying feature synergy, it eliminates isolated abnormal signals caused by equipment errors or environmental interference, accurately locating multi-feature synchronous abnormal areas truly caused by interlayer gaps, thus improving the reliability of the detection results. Furthermore, this method is adaptable to different road segment conditions, requiring no parameter adjustment for specific scenarios, and possesses good versatility.

[0146] This application also provides a device for detecting interlayer voids in cement concrete pavement.

[0147] like Figure 2As shown, the interlayer void detection device 200 for cement concrete pavement includes an acquisition module 210, a first determination module 220, a second determination module 230, a third determination module 240, and a fourth determination module 250. The acquisition module 210 acquires first radar local features and first vibration local features of a first region of the cement concrete pavement, second radar local features and second vibration local features of a second region, and third radar local features and third vibration local features of a third region. The cement concrete pavement is divided into at least three regions, with the first region adjacent to the second and third regions. The first determination module 220 uses the first radar local features, second radar local features, and third radar local features to determine the first local radar feature abrupt change coefficient. The second determination module 230 uses the first vibration local features, second vibration local features, and third vibration local features to determine the first local vibration energy gradient. The third determination module 240 uses the first radar local features and first vibration local features to determine the first synergy difference of the first region. The fourth determination module 250 uses the first synergy difference, local vibration energy gradient and local radar feature mutation coefficient to determine the first local void anomaly degree of the first region, and then determines the interlayer void situation of cement concrete pavement.

[0148] This application effectively solves the problem of inaccurate localization of voided areas in existing detection devices. Through multi-module collaborative processing, it achieves simultaneous detection of radar signal abrupt changes, vibration energy attenuation, and characteristic anomalies, significantly reducing the probability of misjudgment based on a single feature. The device can accurately identify the boundary transition zone between voided and normal areas, avoiding erroneous judgments caused by local material inhomogeneity or environmental interference. By dynamically adjusting global weights, the device can adapt to differences in risk levels across different road sections, ensuring that the detection results match the actual severity of voiding. The final output of local voiding anomaly degree provides a quantitative basis for maintenance decisions, supporting the precise formulation of repair plans.

[0149] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0150] The various embodiments in this specification 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.

Claims

1. A method for detecting interlayer voids in cement concrete pavement, characterized in that, include: The method acquires a first radar local feature and a first vibration local feature of a first region of a cement concrete pavement, a second radar local feature and a second vibration local feature of a second region, and a third radar local feature and a third vibration local feature of a third region. The cement concrete pavement is divided into at least three regions, and the first region is adjacent to the second region and the third region, respectively. The first local radar feature abrupt change coefficient of the first region is determined by using the first radar local feature, the second radar local feature, and the third radar local feature; The first local vibration energy gradient of the first region is determined using the first local vibration feature, the second local vibration feature, and the third local vibration feature. Using the first radar local features and the first vibration local features, the first coordination difference in the first region is determined; Using the first synergy difference, the first local vibration energy gradient, and the first local radar feature mutation coefficient, the first local void anomaly degree of the first region is determined, thereby determining the interlayer void situation of the cement concrete pavement. The method for obtaining the first local radar feature mutation coefficient is as follows: obtaining the first variation amplitude between the first radar local feature and the second radar local feature, and obtaining the second variation amplitude between the first radar local feature and the third radar local feature; using the first variation amplitude and the second variation amplitude, determining the first local radar feature mutation coefficient; obtaining the global voiding anomaly feature of the cement concrete pavement, and determining the global normal fluctuation threshold of the adjacent bimodal difference of the radar frame; using the global normal fluctuation threshold and the maximum value of the first variation amplitude and the second variation amplitude, determining the first local radar feature mutation coefficient. The method for obtaining the first local vibration energy gradient is as follows: using the first vibration local feature and the second vibration local feature, a first difference is determined; using the first vibration local feature and the third vibration local feature, a second difference is determined; using the first difference and the second difference, a directional consistency coefficient is determined; in response to the first vibration local feature being greater than the second vibration local feature, the first local vibration energy gradient is determined using the directional consistency coefficient, the second difference, and the first length of the first region. The method for obtaining the first coordination difference is as follows: using the first radar local features and the first vibration local features, a first coordination coefficient of the first region is determined; using the first coordination coefficient, a coordination average value of all regions of the cement concrete pavement is determined; using the first coordination coefficient and the coordination average value, a third difference value is determined; using the third difference value and the coordination average value, the first coordination difference value is determined; the average value of radar local features and the average value of vibration local features corresponding to all regions of the cement concrete pavement are obtained; using the average value of radar local features and the first radar local features, a first relative anomaly amplitude is determined; and using the average value of vibration local features and the first vibration local features, a second relative anomaly amplitude is determined; using the first relative anomaly amplitude, the second relative anomaly amplitude, and a second preset value, a first matching degree is determined; using the first matching degree, the first relative anomaly amplitude, and the second relative anomaly amplitude, a first coordination coefficient of the first region is determined. The method for obtaining the first local void anomaly degree is as follows: acquiring the global radar void feature and global vibration stiffness void feature of the cement concrete pavement; determining the global void anomaly feature using the global radar void feature and the global vibration stiffness void feature, and determining the global weight using the global void anomaly feature; determining the magnitude and component sum of the void feature vector using the first synergy difference, the first local vibration energy gradient, and the first local radar feature mutation coefficient; and determining the first local void anomaly degree of the first region using the magnitude, the component sum, and the global weight.

2. The method for detecting interlayer voids in cement concrete pavement according to claim 1, characterized in that, The acquisition of the first radar local features and the first vibration local features of the first region of the cement concrete pavement includes: The first mean distance between the first two peaks and the first quantization range of the radar reflection signal of the radar frame in the first region are obtained, and the first mean distance between the first two peaks and the first quantization range are used to determine the normalized value of the first interval between the two peaks. The average value of the first bimodal height difference and the average value of the first maximum peak height of the radar frames in the first region are obtained, and the normalized value of the first bimodal height is determined using the average value of the first bimodal height difference and the average value of the first maximum peak height. The local features of the first radar are determined using the normalized value of the first bimodal spacing and the normalized value of the first bimodal height. Obtain the sum of the first frequency band energy of the low-frequency energy in the first region, and obtain the sum of the second frequency band energy of the full frequency band energy in the first region; The first vibration local characteristics are determined by using the sum of energy in the first frequency band and the sum of energy in the second frequency band.

3. The method for detecting interlayer voids in cement concrete pavement according to claim 2, characterized in that, The step of determining the local features of the first radar using the normalized value of the first bimodal spacing and the normalized value of the first bimodal height includes: The positive correlation index is determined by using the difference between the first preset value and the normalized value of the bimodal height; The local features of the first radar are determined using the positive correlation index and the normalized value of the bimodal spacing. The step of determining the first local vibration characteristics using the sum of energy in the first frequency band and the sum of energy in the second frequency band includes: Obtain the total number of sampling points for the first vibration local feature in the first region; The first vibration local characteristics are determined by using the total number of sampling points, the sum of energy in the first frequency band, and the sum of energy in the second frequency band.

4. A device for detecting interlayer voids in cement concrete pavement, characterized in that, This device is used to implement the interlayer void detection method for cement concrete pavement according to any one of claims 1-3, comprising: The acquisition module is used to acquire the first radar local features and the first vibration local features of the first region of the cement concrete pavement, as well as the second radar local features and the second vibration local features of the second region, and the third radar local features and the third vibration local features of the third region. The cement concrete pavement is divided into at least three regions, and the first region is adjacent to the second region and the third region, respectively. The first determining module uses the first radar local features, the second radar local features, and the third radar local features to determine the first local radar feature abrupt change coefficient; The second determining module uses the first local vibration feature, the second local vibration feature, and the third local vibration feature to determine the first local vibration energy gradient; The third determining module uses the first radar local features and the first vibration local features to determine the first coordination difference in the first region; The fourth determining module uses the first synergy difference, the first local vibration energy gradient, and the first local radar feature mutation coefficient to determine the first local void anomaly degree in the first region, thereby determining the interlayer void situation of the cement concrete pavement.

Images