An ultrasonic reflection method for detecting the depth of cracks in asphalt pavement

Abstract

This invention belongs to the field of pavement crack detection technology, specifically disclosing an ultrasonic reflection detection method for asphalt pavement crack depth. By extracting the crack centerline and feature points to generate a detection cross-section sequence perpendicular to the centerline, the measuring points accurately cover the locations of crack morphological changes, avoiding the loss of depth information caused by equidistant point placement. Furthermore, probe axes are arranged perpendicular to the local crack direction on each detection cross-section to ensure that the ultrasonic diffraction wave propagation path is consistent with the theoretical model, reducing inversion errors introduced by cross-section orientation deviations. Simultaneously, waveform similarity calculations are performed on the characteristic waveform sequences of adjacent cross-sections along the crack direction to identify waveform abrupt change points. These abrupt change points are then used as boundaries to divide the continuous cross-section into several crack segments, allowing the crack depth distribution to be segmented based on the actual crack morphological changes. The final output crack depth distribution truly reflects the development state of the crack along its direction.

Description

Technical Field

[0001] This invention belongs to the field of pavement crack detection technology, and specifically discloses an ultrasonic reflection detection method for the depth of asphalt pavement cracks. Background Technology

[0002] During long-term service, asphalt pavements are prone to cracking due to factors such as temperature and vehicle loads. Crack depth is a key indicator for evaluating the degree of damage to the pavement structure, and accurate detection of crack depth is valuable for preventing further damage to the pavement structure.

[0003] Currently, ultrasonic reflection methods are widely used for detecting the depth of cracks in asphalt pavements due to their advantages of being non-destructive, rapid, and portable. For example, Chinese invention patent CN109839439B discloses a road and bridge pavement crack detection system and its detection method. This system involves installing several ultrasonic transmitting modules and corresponding receiving modules side-by-side on a mobile device, vertically transmitting ultrasonic signals towards the pavement, receiving the reflected echoes, calculating the depth value at each measuring point, and then connecting the coordinates of each reflection point using the position of the ultrasonic transmitting modules as the base axis to draw a two-dimensional cross-sectional view of the crack. Finally, the two-dimensional images from different time points are combined to form a three-dimensional image, ultimately creating a three-dimensional structural image of the crack. This method achieves rapid detection of crack depth over a wide range through the principle of vertical reflection.

[0004] However, the above solutions still have the following technical problems in practical applications: The above-mentioned methods typically deploy measuring points at equal intervals or time intervals along the travel path of the mobile device. This deployment method does not fully consider the local geometric features of the crack, such as its meandering, bifurcating, and misaligned features. When the measuring points do not cover the locations where the crack morphology changes, the position of the ultrasonic probe cannot accurately reflect the actual geometry of the crack, causing the ultrasonic wave propagation path to deviate from the theoretical assumptions, which in turn leads to systematic errors in the depth calculation. Due to the non-uniformity of the pavement structure layer, the crack depth often exhibits an uneven distribution. The above scheme only calculates the depth independently for discrete measuring points. Each measuring point is isolated from the others and lacks continuous analysis. It is difficult to capture the abrupt change position of the crack depth along the direction, resulting in the inability to define the boundary point of crack depth change and insufficient accuracy in characterizing the crack depth distribution characteristics. Summary of the Invention

[0005] To solve the above-mentioned technical problems, or at least partially solve them, this invention provides an ultrasonic reflection detection method for asphalt pavement crack depth based on crack contour guidance and depth abrupt change identification, thereby improving the spatial resolution of crack depth detection.

[0006] The objective of this invention can be achieved through the following technical solution: an ultrasonic reflection detection method for the depth of cracks in asphalt pavement, comprising: The crack centerline and feature points are extracted based on the two-dimensional planar morphology of the crack, and a sequence of detection cross sections perpendicular to the centerline is generated along the crack direction based on the feature points. On each detection section, the transmitting and receiving probes are symmetrically arranged on both sides of the crack. They are moved outward synchronously in a direction perpendicular to the crack direction according to the increasing sequence of probe spacing, and the diffraction waveforms at each probe spacing are collected. Multi-probe spacing waveforms are generated section by section along the detection section sequence. The waveforms of the multi-probe spacing at each cross section are preprocessed and arranged in order of probe spacing to form a characteristic waveform sequence. The waveform similarity of the characteristic waveform sequences of adjacent cross sections is calculated along the crack direction. Based on the waveform similarity of adjacent sections, waveform abrupt change points are identified, and continuous sections are divided into crack segments using the abrupt change points as boundaries. The travel time of diffracted waves is extracted from the waveforms of the multi-probe spacing at each cross-section, and the crack depth value of each cross-section is calculated by combining the preset sound velocity. The crack depth distribution map with the boundary point and the depth value of each cross-section is output.

[0007] Combining all the above technical solutions, the positive effects of this invention are as follows: 1. This invention generates a sequence of detection cross-sections perpendicular to the centerline by extracting the crack centerline and feature points, so that the measuring points accurately cover the locations of crack morphological changes, avoiding the loss of depth information caused by equally spaced points. On this basis, the probe axis is arranged perpendicular to the local direction of the crack in each detection cross-section to ensure that the propagation path of the ultrasonic diffraction wave is consistent with the theoretical model, reducing the inversion error introduced by the cross-section direction deviation and improving the accuracy of depth calculation.

[0008] 2. This invention calculates the waveform similarity of characteristic waveform sequences of adjacent cross sections along the crack direction, identifies waveform abrupt change points, and divides the continuous cross section into several crack segments using the abrupt change points as boundaries. This allows the crack depth distribution to be segmented based on the actual morphological changes of the crack, thereby avoiding depth distortion caused by forcibly merging or smoothing crack segments with different acoustic characteristics. The final output crack depth distribution more realistically reflects the development state of the crack along its direction. Attached Figure Description

[0009] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0010] Figure 1 This is a diagram illustrating the implementation steps of the method of the present invention; Figure 2 This is a flowchart illustrating the implementation process of calculating the waveform similarity of characteristic waveform sequences of adjacent cross sections along the crack direction in this invention. Figure 3 This is a flowchart illustrating the implementation process of dividing a continuous cross-section into crack segments using abrupt change points as boundaries in this invention. Detailed Implementation

[0011] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0012] Please see Figure 1 As shown, this invention proposes an ultrasonic reflection detection method for the depth of cracks in asphalt pavement, including Step 1: extracting the crack centerline and feature points based on the two-dimensional planar morphology of the crack, and generating a detection cross-section sequence perpendicular to the centerline along the crack direction using the feature points as a reference.

[0013] During long-term service, asphalt pavements are prone to cracking due to factors such as temperature and vehicle loads. When using ultrasonic reflection to detect crack depth, probes need to be placed on both sides of the crack to collect diffraction waveforms, thus requiring the location of the detection section to be determined beforehand. However, the actual cracks are not straight but exhibit irregular geometric features such as meandering, bifurcating, and misalignment. If these features are ignored and an evenly spaced layout is used, the detection section may deviate from the changing position of the crack, causing the diffraction wave propagation path to be inconsistent with the theoretical model, resulting in systematic errors in depth inversion.

[0014] Therefore, before formally collecting ultrasonic data, this invention first analyzes the two-dimensional planar morphology of the crack. By extracting the crack centerline and feature points, a sequence of detection cross-sections distributed along the crack direction and perpendicular to the centerline is generated based on these feature points, ensuring that each detection cross-section can accurately reflect the local geometric features of the crack.

[0015] In a specific embodiment of the present invention, the process of extracting the crack centerline and feature points is as follows: S11. Obtain a two-dimensional planar image of the crack from the crack image acquisition device. Since there is a difference in grayscale between the crack area and the background road surface area in the original image, in order to separate the crack from the background road surface area, the image is first subjected to binarization segmentation processing to obtain a binarized image of the crack area. This step can effectively eliminate interference from road surface particles, uneven lighting, etc., and highlight the main shape of the crack.

[0016] S12. Given that the crack region in the binarized image is a continuous black pixel block, and its edge is the boundary line between the crack and the road surface, in order to obtain the accurate geometric boundary of the crack, edge detection is performed on the crack region in the binarized image to extract the edge contour lines on both sides of the crack, and obtain the pixel set of the left and right edges of the crack. These edge pixels are used for subsequent width calculation and bifurcation point identification.

[0017] S13. To quantify the crack's orientation and geometric features, a centerline along the crack's extension direction is needed. This requires a skeletonization process for the crack region, where edge pixels are peeled away layer by layer until a single-pixel width of the remaining connected line is obtained; this connected line is the crack's centerline. The centerline serves two purposes: firstly, it provides a spatial positioning reference for the detection section; secondly, it provides a measurement direction perpendicular to the crack's orientation for width calculation. Discrete pixels are extracted from the centerline at equal arc length intervals to obtain the centerline pixel sequence.

[0018] S14. Draw perpendicular lines to the center line at each point along the crack center line. The Euclidean distance between the intersections of this perpendicular line and the left and right edges is the crack width at that point. This yields the crack width value for each pixel on the crack center line. Then, calculate the rate of change of the width value along the crack direction. The rate of change is equal to the width difference between two adjacent points divided by the arc length of the two points along the center line. The sign of the rate of change reflects whether the width is increasing or decreasing. Points where the rate of change changes from positive to negative or from negative to positive and the change amplitude exceeds twice the average change amplitude of the adjacent area are marked as width abrupt change points. The change amplitude refers to the absolute value of the difference between the change amplitude of the current point and the change amplitude of the adjacent points. The average change amplitude of the adjacent area refers to the arithmetic mean of the change amplitudes of several points before and after the current point, such as three points before and after the current point. Width abrupt change points reflect the location where the local width of the crack changes.

[0019] S15. Perform eight-neighborhood connectivity analysis on the contour lines of the crack edges on both sides, and count the number of edge branches in the neighborhood of each edge pixel. Here, a branch refers to different directions extending from the same pixel along the edge direction. When the number of branches is greater than 2, it indicates that there are multiple cracks converging or bifurcating at that location. Mark these pixels as bifurcation points, which reflect the convergence characteristics of the crack network. Since bifurcation often occupies a small area, the center position of the interconnected group of bifurcation points is taken as the final bifurcation point.

[0020] S16. Extract elevation data from both sides of the crack edge. The elevation data is the vertical height of the point on the edge relative to the road reference surface, which can be obtained by three-dimensional laser scanning. Further calculate the elevation difference between the two edges at the same cross-sectional position (i.e., along the vertical direction of the centerline). This difference represents the vertical displacement of the road surface on both sides of the crack, i.e., the misalignment. Mark the midpoint of the continuous section where the elevation difference exceeds the preset misalignment threshold as the misalignment point. The misalignment point reflects the vertical displacement caused by settlement or compression on both sides of the crack. The misalignment threshold can be set to 3mm.

[0021] S17. The two endpoints of the crack are its start and end points, and must be included in the inspection section. Abrupt width changes, bifurcation points, and misalignment points represent locations where the crack morphology undergoes substantial changes. Using these points together as the positioning reference points for the inspection section ensures that each inspection section accurately passes through the locations where the crack's geometric features change, thus avoiding the loss of depth information due to missed sections and providing correct spatial guidance for subsequent section sequence generation and depth inversion.

[0022] After extracting feature points, a sequence of detection cross-sections continuously distributed along the crack direction can be generated based on these feature points, providing spatial positioning basis for subsequent ultrasonic data acquisition. The specific process is as follows: Using the two endpoints of the crack as the starting and ending cross-sectional locations, and the width abrupt change point, bifurcation point, and misalignment point as mandatory locations, a detection cross-section perpendicular to the tangent direction of the crack centerline is generated at each mandatory location. Setting up detection cross-sections at these locations can avoid the loss of depth information due to the omission of key cross-sections. At the same time, the cross-section direction is perpendicular to the local crack direction, which ensures that the line connecting the transmitting and receiving probes is consistent with the theoretical propagation path of diffracted waves, reducing geometric errors.

[0023] Considering that the crack segment between two adjacent mandatory locations may exhibit a slow, continuous change in morphology, if only the two ends are detected while the intermediate region is ignored, the complete depth distribution of the crack within that segment cannot be determined. Therefore, several intermediate cross-sectional locations are inserted along the centerline segment between two adjacent mandatory locations using an equal arc length method. At each intermediate cross-sectional location, a detection cross-section is generated that is perpendicular to the tangent direction of the crack centerline at that point.

[0024] All the detection sections corresponding to the required positions and the detection sections corresponding to the intermediate positions are arranged in order from the starting point to the ending point along the crack direction to form a detection section sequence. The spatial position order of each section on the crack is clarified by the sorting of the detection sections, which provides a unified spatial index for subsequent calculation of the waveform similarity of adjacent sections and output of depth distribution map along the crack direction.

[0025] Step 2: On each detection section, the transmitting probe and receiving probe are symmetrically arranged on both sides of the crack. They are moved outward synchronously in a direction perpendicular to the crack direction according to the increasing sequence of probe spacing, and the diffraction waveforms at each probe spacing are collected. Multi-probe spacing waveforms are generated section by section along the detection section sequence.

[0026] After generating the detection section sequence, ultrasonic probes can be deployed on each section to collect diffracted wave data. However, the key to crack depth detection lies in obtaining the propagation time of the ultrasonic wave from the transmitting probe, through the crack tip, to the receiving probe. This time is related to the distance between the transmitting and receiving probes, i.e., the probe spacing. If only a single fixed probe spacing is used, the diffracted wave overlaps with the direct wave when the probe spacing is too small, and the signal attenuation is severe when the probe spacing is too large, making it difficult to ensure the distinguishability of the diffracted wave signal under cracks of different depths.

[0027] Therefore, this invention introduces an incremental probe spacing sequence, that is, on each detection section, starting from the initial position close to the edge of the crack, the transmitting probe and the receiving probe are moved outward synchronously step by step to collect a series of diffraction waveforms under different probe spacings, thereby forming a multi-probe spacing waveform for that section.

[0028] In a specific implementation of this invention, the probe spacing increment sequence is determined as follows: The initial probe spacing is taken as the probe spacing when both the transmitting and receiving probes are close to the edge of the crack at the current detection section. The purpose of this is to ensure that the transmitting and receiving probes are both located on the intact road surface on both sides of the crack in the initial position, so as to avoid the probes being suspended or having poor contact.

[0029] Historical measurements of crack depth are extracted from historical detection data of similar cracks along the same road section. These historical measurements are used as the estimated crack depth at the current detection section. The maximum probe spacing is set to several times the estimated depth, for example, three times. The principle for selecting the maximum probe spacing is that, at this probe spacing, the difference between the travel time of the diffracted ultrasonic wave and the travel time of the direct wave should be greater than the time required for the ultrasonic wave to travel back and forth once in the direction of the estimated depth. This is because the difference between the travel time of the diffracted wave and the travel time of the direct wave directly reflects the crack depth information. If the difference is too small, the diffracted wave is easily masked by the direct wave, making it difficult to accurately extract the travel time. By setting the maximum probe spacing to several times the estimated depth, it can be ensured that the diffracted wave and the direct wave are fully separated in the time domain when collecting data at the farthest point, which facilitates subsequent waveform identification and travel time acquisition.

[0030] Between the initial probe spacing and the maximum probe spacing, the probe spacing increments are increased sequentially according to a fixed probe spacing increment to generate a probe spacing increasing sequence. This sequence ensures that the diffraction waveform acquired on each detection section changes monotonically with the probe spacing, thereby forming a clear time-distance relationship and providing uniformly spaced data points for subsequent time-distance curve fitting or time-difference analysis.

[0031] Step 3: Preprocess the waveforms of the multi-probe spacing for each cross section, arrange them in order of probe spacing to form a characteristic waveform sequence, and calculate the waveform similarity of the characteristic waveform sequences of adjacent cross sections along the crack direction.

[0032] After acquiring waveforms at the multi-probe spacing for each section, the next step is to extract diffraction wave travel time information for depth calculation from the waveform data. However, given the unavoidable interference from factors such as instrument triggering jitter during on-site acquisition, the original waveforms contain out-of-band noise and amplitude fluctuations. Directly using the original waveforms for diffraction wave identification and travel time extraction will lead to identification errors, thus affecting the accuracy of depth inversion. Therefore, preprocessing of the multi-probe spacing waveforms is necessary.

[0033] In an optional embodiment of the present invention, the preprocessing includes the following steps performed in sequence: Bandpass filtering is applied to the waveforms of the multi-probe spacing at each cross-section to filter out noise frequency components that are below the lower limit of the ultrasonic probe center frequency and above the upper limit of the center frequency, so that only the effective components related to ultrasonic wave propagation are retained in the waveform, thereby improving the signal-to-noise ratio of the diffraction wave.

[0034] For each filtered waveform, a first-wave alignment operation is performed. Specifically, the noise waveform when there is no effective signal is pre-acquired to determine the maximum amplitude of the background noise. Then, the first peak or trough in each waveform whose amplitude exceeds the maximum amplitude of the background noise is taken as the time zero point. The time axis of all waveforms is shifted to this zero point. The purpose of this is that the triggering time of the ultrasonic transmitter may have random delays. By first-wave alignment, these time base differences can be eliminated, so that all waveforms have a unified starting reference point on the time axis, thereby ensuring the accuracy of the diffraction wave travel time measurement.

[0035] After the first wave is aligned, all waveforms are normalized to scale the maximum absolute amplitude of each waveform to the same reference amplitude range. For example, the maximum amplitude is uniformly set to 1. This can eliminate the amplitude fluctuation caused by differences in probe coupling pressure or local roughness of the road surface.

[0036] Furthermore, after completing the above preprocessing, each cross-section obtains a multi-probe spacing waveform. These waveforms are spliced ​​together end to end or formed into a sequence according to the ascending order of the probe spacing, thus forming the characteristic waveform sequence of the cross-section.

[0037] Considering the material inhomogeneity of asphalt pavement structural layers, crack depth often exhibits an uneven distribution along its direction, meaning the entire crack may consist of multiple crack segments with different acoustic characteristics. If crack depth is calculated directly based on the preprocessed waveform by extracting the travel time of diffracted waves, ignoring the acoustic differences between segments, different crack morphologies will be treated interchangeably, failing to accurately reflect the actual boundary of crack depth variation.

[0038] Therefore, this invention proposes that if a crack has different crack segments, there must be a boundary point between adjacent crack segments. Within the same crack segment, due to the similarity in crack geometry and filling state, the waveforms of adjacent sections should have a high degree of similarity; however, on both sides of the boundary point, the waveform morphology will change abruptly. This characteristic provides a quantifiable starting point for quantitatively identifying crack segment boundaries. That is, by calculating the similarity between the characteristic waveform sequences of adjacent sections, the location of waveform abrupt changes can be automatically detected, thereby dividing the crack segments.

[0039] Please see Figure 2 As shown, therefore, the present invention calculates the waveform similarity of characteristic waveform sequences of adjacent cross-sections along the crack direction, specifically including the following steps: First, for each pair of adjacent cross sections, the characteristic waveform sequence of the cross section that is arranged first along the crack direction is taken as the reference sequence, and the characteristic waveform sequence of the cross section that is arranged second is taken as the sequence to be matched.

[0040] Secondly, since the alignment of two waveform sequences on the time axis directly affects the similarity judgment, if two waveforms have the same shape but have a time offset, direct point-by-point comparison will produce a large error. Therefore, this invention uses the cross-correlation method to measure the similarity of adjacent cross-sectional feature waveform sequences.

[0041] Specifically: Keeping the time axis of the reference sequence fixed, the sequence to be matched is slid along the time axis relative to the reference sequence from the leftmost end to the rightmost end. At each time sampling point slid, the values ​​of two sampling points at the same time position in the reference sequence and the sequence to be matched after the slide are multiplied. All multiplication results are summed to obtain the cumulative product sum at that sliding position. The total number of sliding positions is equal to the sum of the lengths of the reference sequence and the sequence to be matched, minus one, thus covering all alignment states of the sequence to be matched and the reference sequence from partial overlap to complete overlap and then separation again. The cumulative product sum is the cross-correlation value at the corresponding sliding position. This value is maximum when the two waveforms are completely identical and perfectly aligned; it decreases when there are differences or misalignment.

[0042] Next, iterate through all the above sliding positions and take the maximum value of the accumulated product as the waveform similarity between adjacent cross sections. The sliding position corresponding to this maximum value is the optimal time alignment position for the two feature waveform sequences, at which point the two sequences are most morphologically matched. The magnitude of the maximum value directly reflects the degree of similarity in the waveform morphology of the two cross sections: the larger the value, the more similar the waveforms; the smaller the value, the greater the difference in waveforms.

[0043] Finally, the above calculation is repeated for each pair of adjacent sections along the crack direction to obtain the waveform similarity of each pair of adjacent sections.

[0044] Step 4: Identify waveform abrupt change points based on the waveform similarity of adjacent sections, and divide the continuous section into crack segments using the abrupt change points as boundaries.

[0045] Please see Figure 3 As shown, after obtaining the waveform similarity of each pair of adjacent sections, the waveform abrupt change points can be identified based on the level of similarity, and the continuous section can be divided into several crack segments using the abrupt change points as boundaries. The following is a detailed explanation with reference to specific embodiments: S41. Read the waveform similarity of each pair of adjacent sections sequentially along the crack direction, and compare the similarity with the preset lower limit of similarity. The lower limit of similarity can be obtained by performing histogram statistics on the waveform similarity of all adjacent sections of the entire crack, and taking the similarity value corresponding to the valley between the high value area and the low value area in the histogram as the lower limit of similarity. When the waveform similarity of a pair of adjacent sections is lower than the lower limit of similarity, the position of the adjacent section is marked as a waveform abrupt change point, reflecting that the acoustic characteristics of the cracks on both sides of the position have changed.

[0046] S42. Taking the starting section of the crack as the starting point of the first crack segment, search for the first waveform abrupt change point along the crack direction towards the end point, take the abrupt change point as the end point of the first crack segment, and divide all continuous sections between the starting point and the end point into the first crack segment.

[0047] S43. Using each waveform abrupt change point as the end point of the previous crack segment and the starting point of the next crack segment, the continuous cross section between two adjacent waveform abrupt change points is successively divided into intermediate crack segments. S44. Take the last waveform abrupt change point as the starting point of the last crack segment, and take the crack termination section as the ending point, and divide all continuous sections between the two into the last crack segment.

[0048] In particular, when the waveform similarity of adjacent sections of the entire crack is not lower than the lower limit of similarity, all continuous sections are divided into a unique crack segment without any abrupt change points.

[0049] Step 5: Extract the travel time of the diffracted wave from the waveform of the multi-probe spacing of each section, and calculate the crack depth value of each section in combination with the preset sound velocity. Output a crack depth distribution map marked with the boundary point and the depth value of each section.

[0050] Following the previous steps of generating the cross-section sequence, acquiring waveforms at multiple probe intervals, preprocessing the waveforms, calculating the similarity between adjacent cross-sections, and segmenting the cracks, a continuous sequence of detection cross-sections along the crack direction has been obtained, and the boundary points of each crack segment have been identified. Next, the travel time of the diffracted waves can be extracted using the multi-probe diffraction waveforms acquired on each cross-section. Combined with the sound velocity of the asphalt pavement, the crack depth value of each cross-section can be calculated, and the crack depth distribution can be displayed. The specific implementation is as follows: Step 1: Extract the travel time of the diffracted wave from the waveforms of the multi-probe spacing at each cross-section. The travel time of diffracted waves is a key parameter for calculating crack depth, reflecting the time it takes for an ultrasonic wave to travel from the transmitting probe, diffract around the crack tip, and reach the receiving probe. To accurately obtain this time, diffracted wave identification and travel time extraction are required for the waveforms at multiple probe spacings across each cross-section. Given that when ultrasonic waves propagate through a intact road surface, the direct wave arrives first, while the wave diffracted around the crack tip is delayed, and that the travel time of diffracted waves increases monotonically with increasing probe spacing, the following steps can be taken: The pre-processed multi-probe spacing waveforms of each cross-section are sequentially identified as diffracted waves in ascending order of probe spacing. Specifically, a waveform is identified as a diffracted wave when it simultaneously meets the following three conditions: a) The waveform appears after the direct wave corresponding to the probe spacing, that is, the arrival time of the diffracted wave is later than that of the direct wave; b) The waveform envelope exhibits a single-peak or double-peak shape, reflecting the typical signal characteristics of diffraction at the crack tip; c) The arrival time of the waveform increases monotonically with the increase of the probe spacing, indicating that the larger the probe spacing, the longer the diffraction path.

[0051] From the waveforms of the multi-probe spacing at each cross-section, the waveform segment that simultaneously satisfies the above three conditions is selected as the diffraction wave.

[0052] For the waveform at each probe spacing, the time corresponding to the peak position of the diffraction wave envelope is recorded as the arrival time of the diffraction wave at that probe spacing. This peak position reflects the moment when the energy of the diffraction wave is strongest, and it has good noise resistance as a travel time feature point.

[0053] The difference between the arrival time of the diffracted wave and the arrival time of the direct wave in the same waveform corresponding to the initial probe spacing in each cross section is taken as the travel time characteristic value of the diffracted wave in the corresponding cross section. The initial probe spacing is used because the path difference between the diffracted wave and the direct wave is the smallest at this time, the geometric relationship between the travel time difference and the crack depth is the most direct, and it is least affected by the non-uniformity of the road surface.

[0054] The second step involves calculating the crack depth based on the preset sound velocity, as detailed below: Obtain the initial probe spacing at each cross-section when both the transmitting and receiving probes are in close contact with the crack edge, and take half of this initial probe spacing as the half-spacing. The half-spacing is the distance from the crack center to the probe on one side.

[0055] Multiplying the travel time eigenvalue of the diffracted wave by the sound velocity on the asphalt pavement yields the path difference. This path difference equals the difference between the total propagation path of the diffracted wave (i.e., the propagation path from the transmitting probe to the crack tip and then to the receiving probe) and the propagation path of the direct wave (i.e., the propagation path from the transmitting probe to the receiving probe, which is actually the initial probe spacing). Since the diffracted wave travels an extra distance from the probe to the crack tip and back compared to the direct wave, the path difference is precisely the length of this extra distance.

[0056] Adding the path difference to the initial probe spacing yields the total propagation path length of the diffracted wave. Since the transmitting and receiving probes are symmetrically arranged, the distance from the transmitting probe to the crack tip is equal to the distance from the crack tip to the receiving probe. Therefore, halving the total propagation path length gives the one-way path length of the ultrasonic wave from the transmitting probe to the crack tip.

[0057] When the transmitting and receiving probes are symmetrically arranged on both sides of the crack and both are close to the crack edge, the crack tip is located directly below the crack centerline. In this case, a right-angled triangle is formed by the straight line from the transmitting probe to the crack tip, the perpendicular line from the crack tip upwards to the road surface, and the horizontal line segment on the road surface from the transmitting probe to the crack center. Wherein: The hypotenuse represents the length of a one-way path; One right-angled side is half the gap, which is the horizontal distance from the center of the crack to the transmitting probe; The other right-angled side is the crack depth to be determined.

[0058] According to the Pythagorean theorem, given the hypotenuse and one leg, the length of the other leg can be calculated, and this length is the crack depth of the cross section.

[0059] Step 3: Output a crack depth distribution map marked with boundary points and depth values ​​for each cross section. The boundary points of each crack segment are determined sequentially along the crack direction. The boundary point is the location of the waveform abrupt change between adjacent crack segments. For the first crack segment, its starting endpoint is the starting point of the boundary point; for the last crack segment, its ending endpoint is the ending point of the boundary point.

[0060] Using the crack centerline as the baseline, mark the location of each boundary point and the depth value of each section on the crack depth distribution map.

[0061] For example, the labeling method is as follows: the horizontal axis is the cumulative arc length along the crack direction, and the vertical axis is the depth value; a dividing mark line perpendicular to the baseline is drawn at each boundary point; depth value points are marked at each cross-section location, and adjacent points can be connected to form a depth variation curve.

[0062] The crack depth distribution map above provides a clear view of the continuous trend of crack depth along its direction and the location of abrupt changes, offering a clear spatial quantitative basis for maintenance decisions.

[0063] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0064] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0065] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0066] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0067] Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for detecting the depth of cracks in asphalt pavement using ultrasonic reflection, characterized in that, include: The crack centerline and feature points are extracted based on the two-dimensional planar morphology of the crack, and a sequence of detection cross sections perpendicular to the centerline is generated along the crack direction based on the feature points. On each detection section, the transmitting and receiving probes are symmetrically arranged on both sides of the crack. They are moved outward synchronously in a direction perpendicular to the crack direction according to the increasing sequence of probe spacing, and the diffraction waveforms at each probe spacing are collected. Multi-probe spacing waveforms are generated section by section along the detection section sequence. The waveforms of the multi-probe spacing at each cross section are preprocessed and arranged in order of probe spacing to form a characteristic waveform sequence. The waveform similarity of the characteristic waveform sequences of adjacent cross sections is calculated along the crack direction. Based on the waveform similarity of adjacent sections, waveform abrupt change points are identified, and continuous sections are divided into crack segments using the abrupt change points as boundaries. The travel time of diffracted waves is extracted from the waveforms of the multi-probe spacing at each cross-section, and the crack depth value of each cross-section is calculated by combining the preset sound velocity. The crack depth distribution map with the boundary point and the depth value of each cross-section is output.

2. The ultrasonic reflection detection method for asphalt pavement crack depth as described in claim 1, characterized in that: The extraction of the crack centerline and feature points includes the following: Binarize the two-dimensional planar image of the crack to separate the crack region from the background; Edge detection is performed on the crack region in the binarized image to extract the edge contour lines on both sides of the crack, thus obtaining the pixel set of the left and right edges of the crack. Skeletonization is performed on the crack region to obtain a center line with a single pixel width, and the discrete pixel sequence on the center line is extracted. Calculate the crack width value and its rate of change along the crack direction for each pixel along the center line. Mark the points where the rate of change changes from positive to negative or from negative to positive and the change amplitude exceeds twice the average change amplitude of the adjacent area as width abrupt points. Eight-neighborhood connectivity analysis was performed on the contour lines of the crack edges on both sides. The number of edge branches in the neighborhood of each edge pixel was counted. Pixels with more than 2 branches were marked as bifurcation points, and the center of the interconnected bifurcation point group was taken as the final bifurcation point. Elevation data are extracted from both sides of the crack, and the elevation difference between the two sides of the same cross section is calculated. The midpoint of the continuous section where the elevation difference exceeds the preset misalignment threshold is marked as the misalignment point. The two ends of the crack, the point of abrupt change in width, the bifurcation point, and the misalignment point are used together as the positioning reference points for the detection section.

3. The ultrasonic reflection detection method for asphalt pavement crack depth as described in claim 2, characterized in that: The generation of a detection cross-section sequence perpendicular to the centerline along the crack direction, based on feature points, includes the following: Using the two ends of the crack as the starting and ending section positions, and the width change point, bifurcation point, and misalignment point as the mandatory positions, a detection section perpendicular to the tangent direction of the crack centerline at each mandatory position is generated. Intermediate cross-section positions are inserted between adjacent mandatory positions with equal arc lengths, and a detection cross-section perpendicular to the tangent direction of the crack centerline at each intermediate cross-section position is generated. All test sections are arranged in order along the direction of the crack to form a test section sequence.

4. The ultrasonic reflection detection method for the depth of cracks in asphalt pavement as described in claim 1, characterized in that: The increasing sequence of probe spacing is determined according to the following process: The initial probe spacing is taken as the probe spacing when both the transmitting and receiving probes are in close contact with the crack edge at the current detection section. Historical depth values ​​are extracted from historical detection data of the same type of crack in the same road section as the estimated depth, and the maximum probe spacing is several times the estimated depth. Between the initial probe spacing and the maximum probe spacing, the probe spacing increments are increased sequentially according to a fixed probe spacing increment to generate a probe spacing increment sequence.

5. The ultrasonic reflection detection method for asphalt pavement crack depth as described in claim 1, characterized in that: The preprocessing of the multi-probe spacing waveform for each cross-section includes the following steps: Bandpass filtering was applied to the waveforms of the multi-probe spacing at each cross section. Pre-acquire noise waveforms when there is no effective signal to determine the maximum amplitude of background noise; For each filtered waveform, the first peak or trough in each waveform whose amplitude exceeds the maximum amplitude of the background noise is taken as the zero point of time, and the time axis of all waveforms is shifted to that zero point. After the first wave is aligned, all waveforms are normalized to scale the maximum absolute amplitude of each waveform to the same reference amplitude range.

6. The ultrasonic reflection detection method for asphalt pavement crack depth as described in claim 1, characterized in that: The waveform similarity calculation of the characteristic waveform sequences of adjacent cross sections along the crack direction includes the following steps: For each pair of adjacent cross sections, the characteristic waveform sequence of the cross section arranged first along the crack direction is used as the reference sequence, and the characteristic waveform sequence of the cross section arranged second is used as the sequence to be matched. Keeping the time axis of the reference sequence fixed, the sequence to be matched is slid from the leftmost end to the rightmost end along the time axis relative to the reference sequence. For each time sampling point slid, the values ​​of two sampling points at the same time position in the reference sequence and the sequence to be matched after sliding are multiplied, and all the multiplication results are accumulated to obtain the sum of the accumulated products at the corresponding sliding position. Iterate through all sliding positions and take the maximum value among all accumulated product sums as the waveform similarity between adjacent cross sections; The above calculation is repeated for each pair of adjacent cross sections along the crack direction to obtain the waveform similarity of each pair of adjacent cross sections.

7. The ultrasonic reflection detection method for asphalt pavement crack depth as described in claim 6, characterized in that: The step of identifying waveform abrupt change points based on the waveform similarity of adjacent cross sections, and dividing continuous cross sections into crack segments using abrupt change points as boundaries, includes the following: The waveform similarity of each pair of adjacent sections is read sequentially along the crack direction. When the waveform similarity of a pair of adjacent sections is lower than the lower limit of similarity, the position of the adjacent section is marked as a waveform abrupt change point. Using the crack initiation section as the starting point of the first crack segment, search for the first waveform abrupt change point along the crack direction towards the end point, and take this abrupt change point as the end point of the first crack segment. Divide all continuous sections between the starting point and the end point into the first crack segment. By taking each waveform abrupt change point as the end point of the previous crack segment and the starting point of the next crack segment, the continuous cross section between adjacent waveform abrupt change points is divided into intermediate crack segments. Taking the last waveform abrupt change point as the starting point and the crack termination section as the ending point, all continuous sections between the two are divided into the last crack segment.

8. The ultrasonic reflection detection method for the depth of cracks in asphalt pavement as described in claim 1, characterized in that: The extraction of diffracted wave travel time from the waveforms of multiple probes at various cross-sections includes the following: From the pre-processed multi-probe spacing waveforms of each cross-section, waveform segments that simultaneously meet the following three conditions are extracted as diffracted waves: a) The waveform appears after the direct wave corresponding to the probe spacing; b) The waveform envelope exhibits a single-peak or double-peak shape; c) The arrival time of the waveform increases monotonically with the increase of the probe spacing; For the waveform at each probe spacing, the time corresponding to the peak position of the diffraction wave envelope is recorded as the arrival time of the diffraction wave at that probe spacing; The difference between the arrival time of the diffracted wave corresponding to the initial probe spacing in each cross section and the arrival time of the direct wave in the same waveform is taken as the travel time characteristic value of the diffracted wave in the corresponding cross section.

9. The ultrasonic reflection detection method for the depth of cracks in asphalt pavement as described in claim 1, characterized in that: The calculation of crack depth values ​​for each cross-section based on the preset sound velocity is as follows: Obtain the initial probe spacing at each cross section when the transmitting and receiving probes are in close contact with the crack edge, and take half of the initial probe spacing as the half spacing; Multiplying the travel time characteristic value of the diffracted wave by the sound velocity of the asphalt pavement yields the path difference between the propagation path of the diffracted wave and the propagation path of the direct wave. Add the path difference to the initial probe spacing and take half to obtain the one-way path length of the ultrasonic wave from the transmitting probe to the tip of the crack. Using the length of the single-path segment as the hypotenuse and half the distance as one leg, the length of the other leg is calculated using the geometric relationship of a right triangle. This length is the crack depth value.

10. The ultrasonic reflection detection method for the depth of cracks in asphalt pavement as described in claim 1, characterized in that: The crack depth distribution map is generated according to the following process: The boundary points of each crack segment are determined sequentially along the crack direction. The boundary points are the locations of waveform abrupt changes between adjacent crack segments. Using the crack centerline as the baseline, mark the location of each boundary point and the depth of each section on the crack depth distribution map.

Images