A road linear engineering surface crack width distribution detection method, system, device and medium

Abstract

This application provides a distributed detection method, system, equipment, and medium for crack width on linear road engineering surfaces, relating to the field of road crack detection technology. It involves collecting ambient temperature data, acoustic signals of cracks at the interface between the base layer and the surface layer, and optical images of the cracks. The acoustic signals and optical images are denoised to obtain effective acoustic signals and effective optical images. Multi-physics coupling technology is used to correct the effective acoustic signals and imaging distortion deviations, resulting in calibrated acoustic signals and calibrated optical images. The acoustic feature parameters of the calibrated acoustic signals and the crack edge spacing of the calibrated optical images are correlated and fitted with the sample crack width of a preset standard crack sample to establish a benchmark correspondence. Combining the acoustic feature parameters and crack edge spacing, a transfer learning model is used to perform distributed crack width inversion at the interface between the base layer and the surface layer to obtain the actual crack width, thus improving the accuracy of distributed crack width detection.

Description

Technical Field

[0001] This application relates to the field of road crack detection technology, and in particular to a distributed detection method, system, equipment and medium for crack width on linear engineering surfaces of roads. Background Technology

[0002] During the long-term service of linear road engineering, the junction between the base layer and the surface layer is prone to hidden cracks due to factors such as temperature changes, repeated loads, and material aging. Although these cracks are difficult to identify with the naked eye in the early stages, they will significantly weaken the overall performance of the structure and thus induce more serious pavement diseases.

[0003] Currently, existing solutions employ a fusion technology of high-frequency ultrasound and high-resolution vision to simultaneously acquire optical images of the road surface and ultrasonic echo signals from within the structure. Image processing extracts the surface crack contours, and by combining ultrasonic reflection delay and energy attenuation characteristics, the crack depth and approximate width are indirectly inferred. This aims to improve detection reliability by leveraging complementary multi-source information and reduce misjudgments caused by environmental or material factors affecting single-sensor methods. However, existing solutions have significant shortcomings in practical applications. For example, they lack effective compensation for acoustic characteristic fluctuations caused by medium inhomogeneity and temperature changes during ultrasonic propagation, leading to depth inversion errors; they do not adequately correct for the impact of viewing angle, distortion, and texture interference in optical imaging on crack edge recognition, resulting in poor stability in width calculation; and the multi-source data are simply correlated without physical mechanism-based correction, making it difficult to accurately achieve distributed width detection of hidden cracks. Summary of the Invention

[0004] The purpose of this application is to provide a distributed detection method, system, device and medium for the width of surface cracks in linear road engineering, so as to solve the problem of inaccurate detection of the width of hidden cracks caused by the lack of coordinated correction of acoustic and optical deviations in the prior art.

[0005] To address the aforementioned technical problems, in a first aspect, this application provides a distributed detection method for the width of surface cracks in linear road engineering projects, comprising:

[0006] Collect ambient temperature data, acoustic signals of cracks at the junction of the base and surface layers, and optical images of the cracks of the road to be inspected;

[0007] The acoustic signal and the crack optical image are denoised to obtain an effective acoustic signal and an effective optical image;

[0008] Based on the ambient temperature data, the material mechanical parameters and optical property parameters of the pavement material corresponding to the road to be tested, and the size calibration relationship of the crack optical image, the propagation attenuation deviation of the effective acoustic signal and the imaging distortion deviation of the effective optical image are corrected by multi-physics coupling technology to obtain the calibrated acoustic signal and the calibrated optical image.

[0009] The acoustic feature parameters of the calibration acoustic signal and the crack edge spacing of the calibration optical image are correlated and fitted with the sample crack width of the preset standard crack sample to establish a benchmark correspondence.

[0010] Based on the acoustic feature parameters, the crack edge spacing, and the baseline correspondence, a distributed width inversion of the crack at the junction of the base layer and the surface layer is performed using a transfer learning model to obtain the actual crack width.

[0011] Optionally, based on the ambient temperature data, the material mechanical parameters and optical property parameters of the pavement material corresponding to the road to be tested, and the size calibration relationship of the crack optical image, multi-physics coupling technology is used to correct the propagation attenuation deviation of the effective acoustic signal and the imaging distortion deviation of the effective optical image, to obtain a calibrated acoustic signal and a calibrated optical image, including:

[0012] Based on the ambient temperature data and the material mechanical parameters of the pavement material corresponding to the road to be tested, the first correlation between the propagation attenuation deviation of the effective acoustic signal and the material elastic properties and ambient temperature is determined using multi-physics coupling technology.

[0013] Based on the first correlation, the propagation amplitude change and propagation time offset in the effective acoustic signal are adjusted to obtain the attenuation correction amount of the effective acoustic signal;

[0014] Based on the material mechanical parameters, optical property parameters and ambient temperature data of the pavement material corresponding to the road to be detected, a second correlation is determined between the imaging distortion deviation of the effective optical image and the material optical refractive index, mechanical deformation and ambient temperature.

[0015] Based on the imaging distortion deviation, the second correlation relationship, and the size calibration relationship of the slit optical image, the pixel stretching and edge offset in the effective optical image are adjusted to obtain the distortion correction amount of the effective optical image;

[0016] The effective acoustic signal and the effective optical image are corrected according to the attenuation correction amount and the distortion correction amount, respectively, to obtain the calibrated acoustic signal and the calibrated optical image.

[0017] Optionally, based on the ambient temperature data and the material mechanical parameters of the pavement material corresponding to the road to be tested, a first correlation between the propagation attenuation deviation of the effective acoustic signal and the material elastic properties and ambient temperature is determined using multiphysics coupling technology, including:

[0018] Elastic correlation information is extracted from the material mechanics parameters of the pavement material corresponding to the road to be tested. The elastic correlation information includes the elastic response value and deformation recovery rate of the pavement material under different stresses.

[0019] Based on the ambient temperature data, multiple temperature gradient intervals are divided, and the fluctuation range of elastic response values ​​and the trend of deformation recovery rate within each temperature gradient interval are analyzed to generate target elastic information.

[0020] Attenuation feature data is extracted from the effective acoustic signal, and the attenuation feature data includes the amplitude reduction during signal propagation and the duration extension during signal propagation.

[0021] Using multi-physics coupling technology, the fluctuation amplitude and deformation recovery rate of the elastic response value corresponding to each temperature gradient interval in the target elastic information are matched with the amplitude reduction and duration extension. Combined with the influence of the thermal conductivity information of the pavement material of the road to be tested on temperature transmission, an elastic attenuation correlation table is formed.

[0022] Based on the elastic attenuation correlation table, a first correlation is determined between the propagation attenuation deviation of the effective acoustic signal and the material elastic properties and ambient temperature.

[0023] Optionally, based on the material mechanical parameters, optical property parameters, and ambient temperature data of the pavement material corresponding to the road to be detected, a second correlation is determined between the imaging distortion deviation of the effective optical image and the material's optical refractive index, mechanical deformation, and ambient temperature, including:

[0024] Transmission correlation information is extracted from the optical property parameters of the pavement material of the road to be tested, and shape change information is extracted from the material mechanical parameters of the pavement material of the road to be tested.

[0025] Ambient light intensity data is acquired when optical images of cracks are captured on the road to be detected. Combined with ambient temperature data, the magnitude of change in the light transmission correlation information and the rate of change in the shape change information are analyzed to generate change record information.

[0026] The effective optical image is divided into a crack core region image and a crack periphery region image. The pixel stretching degree of the crack core region image and the edge offset distance of the crack periphery region image are detected under different ambient temperatures and ambient light intensities to generate regional distortion information.

[0027] The amplitude of light transmission change and the rate of shape change in the change record information are matched with the degree of pixel stretching and the edge offset distance in the regional distortion information. Combined with the influence of the thermal conductivity information of the pavement material corresponding to the road under test on temperature transmission and the superposition effect of ambient light intensity on light refraction, a distortion association table is formed.

[0028] Based on the distortion correlation table, a second correlation is determined between the imaging distortion deviation of the effective optical image and the material's optical refractive index, mechanical deformation, and ambient temperature.

[0029] Secondly, this application provides a distributed detection method, system, device, and medium for the width of surface cracks in linear road engineering projects, including:

[0030] The acquisition module is used to acquire ambient temperature data, acoustic signals of cracks at the junction of the base layer and the surface layer, and optical images of cracks in the road to be inspected.

[0031] The processing module is used to perform noise reduction processing on the acoustic signal and the crack optical image to obtain an effective acoustic signal and an effective optical image;

[0032] The correction module is used to correct the propagation attenuation deviation of the effective acoustic signal and the imaging distortion deviation of the effective optical image based on the ambient temperature data, the material mechanical parameters and optical property parameters of the pavement material corresponding to the road to be tested, and the size calibration relationship of the crack optical image, using multi-physics coupling technology, to obtain a calibrated acoustic signal and a calibrated optical image.

[0033] The correlation module is used to correlate and fit the acoustic feature parameters of the calibration acoustic signal and the crack edge spacing of the calibration optical image with the sample crack width of the preset standard crack sample to establish a benchmark correspondence.

[0034] The inversion module is used to perform distributed width inversion of the crack at the junction of the base layer and the surface layer based on the acoustic feature parameters, the crack edge spacing and the reference correspondence, and obtain the actual crack width through a transfer learning model.

[0035] Thirdly, this application provides an electronic device, comprising:

[0036] Memory, used to store computer programs;

[0037] A processor is configured to execute the computer program to implement the steps of a distributed detection method for surface crack width of linear engineering roads as described in the first aspect above.

[0038] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps of the distributed detection method for surface crack width of linear engineering road described in the first aspect above.

[0039] This application provides a distributed detection method for the width of surface cracks in linear road engineering. It sequentially collects ambient temperature, crack acoustic signals, and optical images, removing interference to obtain high-quality multi-source data. Then, combining temperature, material mechanics and optical properties, and image calibration relationships, a multi-physics coupling mechanism is used to correct acoustic propagation attenuation and optical imaging distortion, thereby improving the physical consistency between signals and images. Based on this, the calibrated acoustic feature parameters and optical edge spacing are correlated and fitted with standard samples to construct a benchmark correspondence. Finally, a transfer learning model is used to achieve distributed, high-precision inversion of the crack width at the junction of the base layer and the surface layer, realizing a closed loop from initial perception to physical correction and then to intelligent inference, improving the accuracy and robustness of hidden crack width detection.

[0040] Furthermore, by establishing the first correlation between acoustic attenuation and material elastic properties and ambient temperature, the amplitude variation and time delay offset of the acoustic signal are adjusted accordingly. Based on the material mechanical parameters, optical properties and temperature, a second correlation is constructed, and pixel stretching and edge offset are precisely compensated in combination with image size calibration. Finally, the two types of corrections are applied to acoustic and optical data respectively to complete the collaborative calibration and overcome the problems of signal distortion and image inaccuracy. Attached Figure Description

[0041] To more clearly illustrate the technical solutions of the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 A schematic flowchart illustrating a distributed detection method for surface crack width in linear road engineering provided in this application embodiment;

[0043] Figure 2 A schematic diagram illustrating the generation of calibration acoustic signals and calibration optical images using multiphysics coupling technology, as provided in this application embodiment;

[0044] Figure 3 This is a schematic diagram of a distributed detection system for the width of surface cracks in linear road engineering provided in an embodiment of this application. Detailed Implementation

[0045] To address the problem that existing technologies for detecting hidden cracks at the junction of road base and surface layers suffer from signal distortion, image distortion, and loose correlation of multi-source data due to insufficient consideration of the coupling effects of ambient temperature and material properties on acoustic propagation and optical imaging, making it difficult to achieve high-precision distributed inversion of crack width, this application improves the quality of raw data by acquiring ambient temperature, crack acoustic response, and optical images and performing noise reduction. Then, it introduces the mechanical and optical parameters of the road material, combines temperature and image calibration information, and uses a multi-physics coupling method to jointly correct acoustic attenuation and optical distortion, aligning different modal data on a physical scale. Based on this, it establishes a mapping relationship between acoustic features, optical edge spacing, and crack width using standard samples, and uses a transfer learning model to achieve continuous spatial inversion of hidden crack width, thus overcoming the bottleneck of detection bias and insufficient stability caused by the lack of physical consistency calibration in existing solutions.

[0046] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0047] The core of this application is to provide a distributed detection method for the width of surface cracks in linear road engineering projects, and a flowchart of one specific implementation is shown below. Figure 1 As shown, the method includes:

[0048] Step 101: Collect ambient temperature data, acoustic signals of cracks at the junction of the base layer and the surface layer, and optical images of the cracks of the road to be tested.

[0049] In this step, the crack at the junction of the base layer and the surface layer refers to the gap between the base layer and the surface layer in the road structure caused by material shrinkage, temperature changes, and load.

[0050] In this embodiment, ambient temperature data at different sampling points on the road to be tested are collected by a temperature sensor, and acoustic waves are emitted and received by an acoustic sensor close to the crack at the junction of the base layer and the surface layer to obtain the acoustic signal of the crack. At the same time, the crack area is photographed by a high-definition imaging device to obtain an optical image of the crack, ensuring that the three types of data come from the same detection area, providing basic data for subsequent processing.

[0051] Step 102: Denoise the acoustic signal and the crack optical image to obtain an effective acoustic signal and an effective optical image.

[0052] In this step, the effective acoustic signal refers to the acoustic signal whose frequency is within the preset effective range and which is free from environmental noise interference after the acoustic signal of the crack at the junction of the base layer and the surface layer is denoised.

[0053] An effective optical image refers to an image of the crack area that has been denoised and removed, including irrelevant areas such as shadows and road stains, leaving only the crack area.

[0054] In this embodiment, an effective acoustic signal can be obtained by filtering and retaining signal components whose frequencies are within a preset effective acoustic range, and by eliminating interference from irrelevant noise in the environment. Simultaneously, an effective optical image can be obtained by identifying and removing irrelevant areas such as shadows and road stains from the crack optical image, retaining only the image region containing the crack.

[0055] Step 103: Based on the ambient temperature data, the material mechanical parameters and optical property parameters of the pavement material corresponding to the road to be tested, and the size calibration relationship of the crack optical image, the propagation attenuation deviation of the effective acoustic signal and the imaging distortion deviation of the effective optical image are corrected by multi-physics coupling technology to obtain the calibrated acoustic signal and the calibrated optical image.

[0056] In this step, the size calibration relationship refers to the conversion relationship between the actual physical size of a single pixel in the optical image of the crack.

[0057] Propagation attenuation deviation refers to the deviation in signal amplitude and propagation time caused by changes in ambient temperature and fluctuations in the mechanical parameters of pavement materials when an effective acoustic signal propagates in the crack area at the junction of the base layer and the surface layer.

[0058] Imaging distortion deviation refers to the deviation of effective optical images caused by changes in the optical property parameters of road paving materials due to changes in ambient temperature, material mechanical deformation, and the influence of imaging angle, resulting in pixel stretching and edge offset.

[0059] Step 104: The acoustic characteristic parameters of the calibration acoustic signal and the crack edge spacing of the calibration optical image are correlated and fitted with the sample crack width of the preset standard crack sample to establish a benchmark correspondence.

[0060] In this step, the preset standard crack sample refers to standard road base and surface specimens that include different known sample crack widths.

[0061] The benchmark correspondence refers to the regular relationship between the acoustic characteristic parameters of the calibration acoustic signal, the crack edge spacing of the calibration optical image, and the sample crack width of the preset standard crack sample by correlating and fitting them.

[0062] Step 105: Based on the acoustic feature parameters, the crack edge spacing, and the baseline correspondence, a distributed width inversion is performed on the crack at the junction of the base layer and the surface layer using a transfer learning model to obtain the actual crack width.

[0063] This application provides complete basic data for detection by collecting ambient temperature data, crack acoustic signals, and optical images; ensures data quality by eliminating data interference; improves data accuracy by using multi-physics coupling technology to correct deviations; provides a reliable reference by establishing a benchmark correspondence; and obtains accurate actual crack width by using a transfer learning model to achieve distributed width inversion.

[0064] This application provides a specific embodiment, such as Figure 2 As shown, step 103 involves using the ambient temperature data, the material mechanical parameters and optical property parameters of the pavement material corresponding to the road to be tested, and the size calibration relationship of the crack optical image to correct the propagation attenuation deviation of the effective acoustic signal and the imaging distortion deviation of the effective optical image using multi-physics coupling technology, thereby obtaining a calibrated acoustic signal and a calibrated optical image. Specifically, this includes the following steps:

[0065] Step 301: Based on the ambient temperature data and the material mechanical parameters of the pavement material corresponding to the road to be tested, the first correlation between the propagation attenuation deviation of the effective acoustic signal and the material elastic properties and ambient temperature is determined using multi-physics coupling technology.

[0066] In this step, the first correlation refers to the correspondence between the effective acoustic signal propagation attenuation deviation and the material elastic properties and ambient temperature, established by using multi-physics coupling technology based on ambient temperature data and the mechanical parameters of road paving materials. This first correlation includes the influence of changes in the material elastic properties at different temperatures on the propagation amplitude and propagation time offset.

[0067] Step 302: Based on the first correlation, adjust the propagation amplitude change and propagation time offset in the effective acoustic signal to obtain the attenuation correction amount of the effective acoustic signal.

[0068] In this step, the change in propagation amplitude refers to the increase or decrease in the actual amplitude of the effective acoustic signal compared to the standard amplitude due to changes in the elastic properties of the material and the influence of ambient temperature during propagation.

[0069] Propagation time offset refers to the increase or decrease in the actual propagation time of an effective acoustic signal compared to the standard time due to changes in the elastic properties of materials and the influence of ambient temperature during propagation.

[0070] The attenuation correction amount refers to the specific value calculated based on the first correlation to compensate for changes in the effective acoustic signal propagation amplitude and propagation time offset. This correction amount includes amplitude compensation value and time compensation value.

[0071] In this embodiment, the changes in propagation amplitude and propagation time offset caused by variations in material elastic properties and ambient temperature in the effective acoustic signal are first identified. Then, based on the corresponding patterns of propagation attenuation deviation, material elastic properties, and ambient temperature in the first correlation relationship, an attenuation correction amount is calculated. The attenuation correction amount includes an amplitude compensation value and a time compensation value. The magnitude of the amplitude compensation value is equal to the magnitude of the propagation amplitude change but opposite in sign, and it is used to compensate for the propagation amplitude change in the effective acoustic signal. The magnitude of the time compensation value is equal to the magnitude of the propagation time offset but opposite in sign, and it is used to compensate for the propagation time offset in the effective acoustic signal.

[0072] The amplitude compensation value can be calculated by multiplying the amplitude attenuation coefficient corresponding to a unit change in material elastic properties (preset in the first correlation) by the actual change in material elastic properties, and by multiplying the amplitude attenuation coefficient corresponding to a unit change in ambient temperature (preset in the first correlation) by the actual change in ambient temperature. The actual change in material elastic properties is the difference between the current elastic response value and the standard elastic response value, and the actual change in ambient temperature is the difference between the current ambient temperature and the standard ambient temperature.

[0073] The time compensation value can be calculated by multiplying the time offset coefficient corresponding to the unit change in material elastic properties in the first correlation by the actual change in material elastic properties and the time offset coefficient corresponding to the unit change in ambient temperature in the first correlation by the actual change in ambient temperature.

[0074] Step 303: Based on the material mechanical parameters, optical property parameters and ambient temperature data of the pavement material corresponding to the road to be tested, determine the second correlation between the imaging distortion deviation of the effective optical image and the material optical refractive index, mechanical deformation and ambient temperature.

[0075] In this step, the second correlation refers to the correspondence between effective optical image imaging distortion deviation and material optical refractive index, mechanical deformation, and ambient temperature. This second correlation includes the influence of material optical refractive index and mechanical deformation on pixel stretching and edge offset at different temperatures.

[0076] Step 304: Based on the imaging distortion deviation, the second correlation relationship, and the size calibration relationship of the crack optical image, adjust the pixel stretching and edge offset in the effective optical image to obtain the distortion correction amount of the effective optical image.

[0077] In this step, the distortion correction amount refers to the specific value used to correct pixel stretching and edge offset in the effective optical image. The distortion correction amount includes pixel stretching correction value and edge offset correction value.

[0078] In this embodiment, pixel stretching and edge shift caused by changes in the optical refractive index and mechanical deformation of the material in the effective optical image can be identified. Then, by combining the size calibration relationship of the crack optical image and the corresponding patterns of imaging distortion deviation, material optical refractive index, mechanical deformation, and ambient temperature in the second correlation relationship, an imaging distortion correction amount can be calculated. The imaging distortion correction amount includes a pixel stretching correction value and an edge shift correction value. The pixel stretching correction value is equal in magnitude but opposite in sign to the pixel stretching value, and is used to offset pixel stretching in the effective optical image. Similarly, the edge shift correction value is equal in magnitude but opposite in sign to the edge shift value, and is used to offset edge shift in the effective optical image.

[0079] The pixel stretching correction value can be obtained by multiplying the pixel stretching coefficient corresponding to the unit change in material optical refractive index (preset in the second correlation) by the actual change in material optical refractive index, and summing this sum with the pixel stretching coefficient corresponding to the unit mechanical deformation of the material (preset in the second correlation) multiplied by the actual change in material mechanical deformation. This preliminary pixel stretching correction value is then converted to a pixel stretching correction value for the actual physical dimensions, ensuring a match with the actual dimensions, in conjunction with the dimensional calibration relationship.

[0080] The edge offset correction value can be obtained by multiplying the edge offset coefficient corresponding to the unit change in material optical refractive index in the second correlation by the actual change in material optical refractive index, and summing this result with the edge offset coefficient corresponding to the unit mechanical deformation of material in the second correlation by the actual change in material mechanical deformation. This preliminary edge offset correction value is then converted into an edge offset correction value for the actual physical dimension using the same dimensional calibration relationship.

[0081] Step 305: Correct the effective acoustic signal and the effective optical image according to the attenuation correction amount and the distortion correction amount respectively to obtain the calibrated acoustic signal and the calibrated optical image.

[0082] In this embodiment, the propagation amplitude variation and propagation time offset in the effective acoustic signal are first compensated based on the attenuation correction amount to eliminate the interference of ambient temperature and material elastic properties on the acoustic signal, thereby obtaining a calibrated acoustic signal. Simultaneously, pixel stretching and edge offset in the effective optical image are corrected based on the imaging distortion correction amount. Finally, the image size is adjusted according to the size calibration relationship to match the actual size, resulting in a calibrated optical image.

[0083] Optionally, step 301 involves determining, based on the ambient temperature data and the material mechanical parameters of the pavement material corresponding to the road to be tested, a first correlation between the propagation attenuation deviation of the effective acoustic signal and the material elastic properties and ambient temperature using multiphysics coupling technology. This specifically includes the following steps:

[0084] Step 311: Extract elastic correlation information from the material mechanics parameters of the pavement material corresponding to the road to be tested. The elastic correlation information includes the elastic response value and deformation recovery rate of the pavement material under different stresses.

[0085] In this step, elastic correlation information refers to the set of data extracted from the mechanical parameters of pavement materials that are related to the material's resistance to deformation and stress recovery performance.

[0086] Elastic response value refers to the specific value of elastic deformation of road paving material under different stresses. Its change with ambient temperature directly affects the propagation amplitude of effective acoustic signals.

[0087] In this embodiment of the application, parameters related to the material's resistance to deformation and recovery performance after being subjected to stress are extracted from the material mechanical parameters of the pavement material corresponding to the road to be tested. These parameters specifically include the elastic response value and deformation recovery rate of the pavement material under different stresses, and these parameters are integrated into elastic correlation information.

[0088] Step 312: Based on the ambient temperature data, divide the data into multiple temperature gradient intervals and analyze the fluctuation range of the elastic response value and the trend of deformation recovery rate within each temperature gradient interval to generate target elastic information.

[0089] In this step, the temperature gradient range refers to a continuous temperature range divided according to preset temperature intervals based on ambient temperature data.

[0090] The target elasticity information refers to the material elasticity property analysis results organized according to temperature gradient ranges. This information includes the fluctuation range of elastic response values ​​and the trend of deformation recovery rate within each temperature range.

[0091] In this embodiment, the collected ambient temperature data is first divided into multiple temperature gradient intervals according to preset temperature intervals. Then, for each temperature gradient interval, the fluctuation amplitude of the elastic response value and the changing trend of the deformation recovery rate in the elastic correlation information are analyzed. The fluctuation amplitude of the elastic response value is the ratio of the maximum difference to the minimum value of the elastic response value within the same temperature interval, and the changing trend of the deformation recovery rate is the direction of increase or decrease in the deformation recovery rate as the temperature rises or falls. Finally, these analysis results are organized according to the temperature gradient intervals to generate target elastic information.

[0092] Step 313: Extract attenuation feature data from the effective acoustic signal. The attenuation feature data includes the amplitude reduction during signal propagation and the duration extension during signal propagation.

[0093] In this step, attenuation characteristic data refers to the core data extracted from the effective acoustic signal that reflects the signal propagation attenuation.

[0094] The amplitude reduction during signal propagation refers to the reduction in the actual amplitude of the effective acoustic signal from the transmitter to the receiver compared to the standard amplitude without attenuation.

[0095] The duration extension in the signal propagation path refers to the increase in the actual propagation time of the effective acoustic signal from the transmitter to the receiver compared to the standard duration without offset.

[0096] In this embodiment, the standard propagation amplitude and standard propagation duration of the effective acoustic signal are first determined. These are the signal amplitude and propagation time values ​​when the effective acoustic signal has no propagation attenuation deviation and no propagation time offset, under conditions of crack-free pavement material of the same type as the road to be tested, standard ambient temperature, and standard detection distance. The standard propagation amplitude is the initial peak amplitude collected by the signal receiver under this condition, and the standard propagation duration is the average of the theoretical propagation time from the transmitter to the receiver under this condition and the actual measured propagation time.

[0097] Then, the reduction in amplitude during the actual signal propagation process compared to the standard amplitude is calculated to obtain the amplitude reduction during signal propagation; at the same time, the increase in actual signal propagation time compared to the standard propagation time is calculated to obtain the duration extension in the signal propagation path. The amplitude reduction and duration extension are integrated into attenuation characteristic data.

[0098] Step 314: Using multiphysics coupling technology, the fluctuation amplitude and deformation recovery rate of the elastic response value corresponding to each temperature gradient interval in the target elastic information are matched with the amplitude reduction and duration extension, and combined with the influence of the thermal conductivity information of the pavement material of the road to be tested on temperature transmission, to form an elastic attenuation correlation table.

[0099] In this step, the influence of heat conduction information on temperature transmission refers to the effect of the temperature transmission rate and uniformity within the material on the material properties, as determined by the analysis of the heat conduction information of the pavement material. For example, uneven temperature transmission can lead to differences in local elastic properties of the material, which in turn affects acoustic signal propagation and optical imaging.

[0100] The elastic attenuation correlation table refers to the corresponding table of material elastic properties and acoustic attenuation characteristics organized according to temperature gradient ranges. The correlation table includes the matching results of the fluctuation range of elastic response value, the trend of deformation recovery rate change and the amount of reduction in amplitude and duration in each temperature range, as well as the influence of heat conduction information on temperature transfer.

[0101] In this embodiment, the target elastic information is divided into temperature gradient intervals. The fluctuation amplitude and deformation recovery rate of the elastic response value in each temperature gradient interval are then matched with the amplitude reduction and duration extension within the same temperature gradient interval to obtain matching results. Furthermore, based on the thermal conductivity information of the pavement material corresponding to the road to be tested, the temperature transmission pattern within the material and its impact on acoustic signal propagation are analyzed. The matching results and the impact of temperature transmission are then compiled into a table to form an elastic attenuation correlation table.

[0102] Step 315: Based on the elastic attenuation correlation table, determine the first correlation between the propagation attenuation deviation of the effective acoustic signal and the material elastic properties and ambient temperature.

[0103] In this embodiment of the application, the correlation between the fluctuation amplitude and the reduction in amplitude of elastic response value and the change trend of deformation recovery rate and the extension of time in different temperature gradient intervals in the elastic attenuation correlation table are analyzed. Then, combined with the influence of heat conduction information on temperature transmission, the unified law of the propagation attenuation deviation of effective acoustic signal with the changes of material elastic properties and ambient temperature is summarized, and finally the first correlation is determined.

[0104] For example, taking asphalt pavement testing as an example, assume that the elastic decay correlation table records data for three temperature gradient intervals: in the temperature gradient interval of 0-10℃, the elastic response value fluctuates by ±0.5GPa, the amplitude decreases by 3dB, the deformation recovery rate is 0.3mm / s, and the duration increases by 6ms; in the temperature gradient interval of 11-25℃, the elastic response value fluctuates by ±1.0GPa, the amplitude decreases by 1.5dB, the deformation recovery rate is 0.8mm / s, and the duration increases by 2ms; in the temperature gradient interval of 26-40℃, the elastic response value fluctuates by ±1.5GPa, the amplitude decreases by 2.2dB, the deformation recovery rate is 1.2mm / s, and the duration increases by 3ms.

[0105] Based on the information that asphalt has slow thermal conductivity, it can be found that when the temperature gradient range is 11-25℃, the elastic response value fluctuates moderately with the smallest amplitude reduction, the deformation recovery rate is fast, and the duration extension is minimal. The further the temperature deviates from this range, the greater the amplitude reduction and the greater the duration extension. Finally, the first correlation relationship was determined.

[0106] Optionally, step 303, based on the material mechanical parameters, optical property parameters, and ambient temperature data of the pavement material corresponding to the road to be detected, determines a second correlation between the imaging distortion deviation of the effective optical image and the material optical refractive index, mechanical deformation, and ambient temperature, specifically including the following steps:

[0107] Step 321: Extract light transmission correlation information from the optical property parameters of the pavement material of the road to be tested, and extract shape change information from the material mechanical parameters of the pavement material of the road to be tested.

[0108] In this step, the light transmission correlation information refers to the set of data related to light propagation extracted from the optical property parameters of the pavement material. This light transmission correlation information includes parameters such as the depth of light penetration through the material and the refraction angle of light when passing through the material. Its changes with ambient temperature and light intensity will affect the imaging quality of the effective optical image.

[0109] Shape change information refers to the set of data related to material deformation extracted from the mechanical parameters of road paving materials. This shape change information includes parameters such as the amount of shrinkage and expansion of the material caused by temperature. These changes will directly lead to distortions such as pixel stretching and edge shift in the effective optical image.

[0110] In this embodiment, parameters related to the depth of light penetration and the refraction angle of light passing through the material are extracted from the optical property parameters of the pavement material of the road to be tested, and these parameters are integrated into light transmission correlation information. At the same time, parameters related to the shrinkage and expansion of the material due to temperature are extracted from the material mechanical parameters, and these parameters are integrated into shape change information.

[0111] Step 322: Obtain ambient light intensity data when the optical image of the crack in the road to be detected is collected, and combine it with the ambient temperature data to analyze the change amplitude of the light transmission correlation information and the change rate of the shape change information to generate change record information.

[0112] In this step, ambient light intensity data refers to the light intensity value of the surrounding environment of the road to be detected when acquiring optical images of cracks.

[0113] The change record information refers to the material property change data organized according to the temperature-light intensity combination. This record includes the light transmission change amplitude and shape change rate of light transmission correlation information under different combinations.

[0114] In this embodiment, ambient light intensity data can be collected by a light intensity sensor when collecting optical images of cracks in the road to be detected. This data is correlated with ambient temperature data according to the collection time. Then, the changes in light penetration depth and refraction angle in the light transmission correlation information under different temperatures and ambient light intensities are analyzed to obtain the light transmission change amplitude. In addition, the changes in material shrinkage and expansion in the shape change information are analyzed to obtain the shape change rate. The obtained light transmission change amplitude and shape change rate are organized according to the temperature-light intensity combination to generate change record information.

[0115] Step 323: Divide the effective optical image into a crack core region image and a crack periphery region image, and detect the pixel stretching degree of the crack core region image and the edge offset distance of the crack periphery region image under different ambient temperatures and ambient light intensities to generate regional distortion information.

[0116] In this step, the crack core region image refers to the image region separated from the effective optical image through image segmentation, which includes only the image region of the main part of the crack.

[0117] The image of the area surrounding the crack refers to the image region surrounding the core region of the crack, which is separated from the effective optical image through image segmentation.

[0118] Pixel stretching refers to the horizontal or vertical scaling ratio of pixels in an effective optical image compared to their standard state.

[0119] Edge offset distance refers to the offset of the crack edge in the effective optical image relative to the actual crack edge.

[0120] Regional distortion information refers to effective optical image distortion data organized by temperature-light intensity-region. This regional distortion information includes the pixel stretching degree of the image of the core region of the crack and the edge offset distance of the image of the surrounding region of the crack under different conditions.

[0121] In this embodiment, an image segmentation algorithm can be used to divide the effective optical image into two parts: an image of the crack core region and an image of the crack periphery surrounding the crack core region. Then, under different ambient temperature and light intensity conditions, the horizontal / vertical scaling ratio of pixels in the crack core region image is measured to obtain the pixel stretching degree, and the distance between the crack edge and the actual crack edge in the crack periphery image is obtained to obtain the edge offset distance. These measurement results are then processed according to temperature-light intensity-region to generate regional distortion information.

[0122] Step 324: Match the light transmission change amplitude and shape change rate in the change record information with the pixel stretching degree and edge offset distance in the regional distortion information, and combine the influence of the thermal conduction information of the pavement material corresponding to the road to be detected on temperature transmission and the superposition effect of ambient light intensity on light refraction to form a distortion association table.

[0123] In this step, the amplitude of light transmission change refers to the specific increase or decrease in the depth of light penetration and the angle of refraction as a function of ambient temperature and light intensity in the light transmission-related information.

[0124] The rate of shape change refers to the rate at which the shrinkage and expansion of a material change with the ambient temperature.

[0125] The superposition effect of ambient light intensity on light refraction refers to the additional influence of changes in ambient light intensity on the refraction angle of light as it passes through pavement materials. For example, increased light intensity will increase the refraction angle, thereby exacerbating pixel stretching of the effective optical image.

[0126] The distortion correlation table refers to a table that organizes material properties and image distortion according to the combination of temperature and light intensity. The distortion correlation table includes the matching results of light transmission change amplitude, shape change rate and pixel stretching degree, edge offset distance under each combination, as well as the influence of heat conduction information and light intensity superposition effect.

[0127] In this embodiment, the change record information is first broken down into temperature-light intensity combinations, and the transmittance change amplitude and shape change rate under each temperature and light intensity combination are extracted. Then, these are matched with the pixel stretching degree and edge offset distance under the same temperature and light intensity combination to obtain matching results. Simultaneously, combined with the thermal conductivity information of the pavement material corresponding to the road to be detected, the influence of temperature transfer on material deformation and the superposition effect of ambient light intensity on light refraction are analyzed to obtain influence analysis results. The matching results and influence analysis results are then organized into a tabular form to form a distortion correlation table.

[0128] Step 325: Based on the distortion correlation table, determine the second correlation between the imaging distortion deviation of the effective optical image and the material optical refractive index, mechanical deformation, and ambient temperature.

[0129] In this embodiment, the correspondence between the transmittance change amplitude and pixel stretching degree, shape change rate and edge offset distance under different temperature-light intensity combinations is analyzed. At the same time, the influence of heat conduction information on temperature transmission and the superposition effect of ambient light intensity on light refraction are combined to summarize the unified law of effective optical image imaging distortion deviation with material optical refractive index, mechanical deformation and ambient temperature changes, and finally determine the second correlation relationship.

[0130] For example, taking asphalt pavement inspection as an example, assume that the distortion correlation table records 3 sets of temperature-light intensity combination data: in the 10℃-5000lux combination data, the light transmittance change is 10%, the pixel stretching is 3%, the shape change rate is 0.1mm / h, and the edge offset distance is 0.2mm; in the 25℃-8000lux combination data, the light transmittance change is 5%, the pixel stretching is 1%, the shape change rate is 0.3mm / h, and the edge offset distance is 0.1mm; in the 40℃-12000lux combination data, the light transmittance change is 8%, the pixel stretching is 2.5%, the shape change rate is 0.5mm / h, and the edge offset distance is 0.3mm.

[0131] Combining the effects of slow thermal conductivity of asphalt and increased ambient light intensity exacerbating light refraction, we can obtain the following: under the 25℃-8000 lux combination, the light transmittance change is small, the pixel stretching is the lowest, the shape change rate is moderate, and the edge offset distance is the smallest. The further the temperature deviates from 25℃ or the light intensity deviates from 8000 lux, the more obvious the pixel stretching and edge offset distance become. Finally, the second correlation relationship was determined.

[0132] The embodiments of this application accurately compensate for the propagation deviation of the effective acoustic signal by calculating the attenuation correction amount; and fully correct the distortion of the effective optical image by calculating the imaging distortion correction amount.

[0133] This application provides a specific embodiment. Step 104 involves correlating and fitting the acoustic characteristic parameters of the calibration acoustic signal and the crack edge spacing of the calibration optical image with the sample crack width of a preset standard crack sample to establish a benchmark correspondence. This specifically includes the following steps:

[0134] Step 401: Based on the ambient temperature data, construct a simulated detection environment for the road to be detected, and collect standard acoustic signals and standard optical images of preset standard crack samples in the simulated detection environment.

[0135] In this step, the simulated testing environment refers to an experimental environment that is consistent with the actual testing conditions, constructed using environmental simulation equipment based on the ambient temperature data of the road to be tested.

[0136] Standard acoustic signals refer to acoustic signals obtained by acquiring preset standard crack samples through acoustic sensors in a simulated testing environment.

[0137] Standard crack optical images refer to optical images obtained by taking pictures of preset standard crack samples in a simulated testing environment using high-definition imaging equipment. They clearly present the morphology and edge information of the cracks in the samples.

[0138] In this embodiment, the road surface testing environment under the ambient temperature conditions of the road to be tested can be reconstructed using an environmental simulation device based on the ambient temperature data of the road to be tested, thus constructing a simulated testing environment consistent with the actual working conditions of the road to be tested. Then, preset standard crack samples with the same type of pavement material as the road to be tested and including different known sample crack widths are selected and placed in the simulated testing environment. Standard acoustic signals of these samples are collected using acoustic sensors, and standard crack optical images of these samples are captured using a high-definition imaging device.

[0139] Step 402: Denoise the standard acoustic signal and the standard crack optical image to obtain the standard effective acoustic signal and the standard effective optical image.

[0140] In this embodiment, components within a preset effective frequency range of the standard acoustic signal are selected and retained, while environmental noise interference is removed to obtain a standard effective acoustic signal. Simultaneously, irrelevant areas such as shadows and stains are removed from the standard crack optical image, retaining only the image region containing the crack, thus obtaining a standard effective optical image.

[0141] Step 403: Based on the ambient temperature data, the material mechanical parameters and optical property parameters of the preset standard crack sample, and the size calibration relationship of the standard crack optical image, the propagation attenuation deviation of the standard effective acoustic signal and the imaging distortion deviation of the standard effective optical image are corrected to obtain the standard calibrated acoustic signal and the standard calibrated optical image.

[0142] In this step, the standard calibration acoustic signal refers to the acoustic signal obtained after correcting the propagation attenuation deviation of the standard valid acoustic signal, which eliminates the interference of ambient temperature and material properties on signal propagation.

[0143] A standard calibrated optical image is an image obtained by correcting the imaging distortion deviation of a standard valid optical image, thus eliminating the interference of ambient temperature and material properties on the imaging.

[0144] In this embodiment, the propagation attenuation deviation of the standard effective acoustic signal is first analyzed and compensated, and then the imaging distortion deviation of the standard effective optical image is analyzed and corrected, finally obtaining the standard calibrated acoustic signal and standard calibrated optical image with interference and deviation eliminated.

[0145] Step 404: Extract standard acoustic feature parameters from the standard calibration acoustic signal. The standard acoustic feature parameters include the peak value of the signal amplitude and the signal propagation time. The pixel spacing between the two edges of the crack in the standard calibration optical image is used as the standard crack edge spacing.

[0146] In this step, the standard acoustic characteristic parameter refers to the acoustic index extracted from the standard calibration acoustic signal that reflects the width of the sample crack. This parameter includes the peak value of the signal amplitude and the signal propagation time.

[0147] The edges on both sides of the crack refer to the boundary contour lines between the crack body and the surrounding material in the standard calibrated optical image.

[0148] The standard crack edge spacing refers to the pixel spacing between the two edges of a crack in a standard calibrated optical image, measured according to the size calibration relationship of the standard crack optical image.

[0149] In this embodiment, the standard calibration acoustic signal is first analyzed in the time domain to identify the maximum amplitude of the waveform as the signal amplitude peak value. The time it takes for the standard calibration acoustic signal to propagate from the transmitter to the receiver is recorded as the signal propagation duration. These two parameters are then integrated into standard acoustic characteristic parameters. Next, based on the size calibration relationship of the standard crack optical image, the number of pixels between the two edges of the crack in the standard calibration optical image is measured, and the spacing between these pixels is taken as the standard crack edge spacing.

[0150] Step 405: Based on the type of road paving material and the ambient temperature data, analyze the changing trend of the standard acoustic characteristic parameters, the correlation between the standard crack edge spacing and the sample crack width of the preset standard crack sample, in order to establish a benchmark correspondence.

[0151] In this step, the correlation refers to the correspondence between standard acoustic characteristic parameters, standard crack edge spacing, and sample crack width of preset standard crack samples under specific pavement material types and ambient temperature conditions.

[0152] In this embodiment, for the same type of pavement material, different ambient temperature data can be combined to analyze the changing trends of standard acoustic characteristic parameters with sample crack width, and the correspondence between standard crack edge spacing and sample crack width. Then, through data fitting, the inherent correlation between standard acoustic characteristic parameters, standard crack edge spacing, and sample crack width is established, forming a benchmark correspondence.

[0153] The embodiments of this application overcome the problems of acoustic and optical signal distortion, edge recognition deviation, and shallow fusion of multi-source information under environmental interference, thereby improving the accuracy and adaptability of hidden crack width detection.

[0154] This application provides a specific embodiment. Step 105 involves using a transfer learning model to perform distributed width inversion on the crack at the junction of the base layer and the surface layer based on the acoustic feature parameters, the crack edge spacing, and the reference correspondence, to obtain the actual crack width. This specifically includes the following steps:

[0155] Step 501: Based on the temperature adaptation range and material adaptation parameters in the benchmark correspondence, the acoustic feature parameters and the crack edge spacing are weighted and integrated to form an input dataset, and the corresponding material type label and temperature label are marked in the input dataset.

[0156] In this step, the temperature adaptation range refers to the continuous range divided according to the ambient temperature in the reference correspondence. This temperature adaptation range is determined based on the material properties and signal propagation laws at different temperatures.

[0157] Material compatibility parameters refer to the weighting coefficients and parameter adjustment values ​​in the benchmark correspondence that match a specific type of pavement material.

[0158] The input dataset refers to the standardized dataset formed by weighting and integrating acoustic feature parameters and crack edge spacing according to material adaptation parameters.

[0159] Material type labeling is used to identify the type of pavement material in the input dataset and the parameter transformation dataset.

[0160] Temperature labels are used to indicate the range of ambient temperature in the input dataset and the parameter transformation dataset.

[0161] In this embodiment, the temperature adaptation range and material adaptation parameters are first extracted from the baseline correspondence. The weighting of acoustic characteristic parameters and crack edge spacing is then determined based on the material adaptation parameters. Finally, the two are weighted and summed to form an input dataset, and the corresponding material type and temperature labels are annotated in the input dataset.

[0162] Step 502: Use a transfer learning model to transform the input dataset into a parameter transformation dataset.

[0163] In this step, the parameter transformation dataset refers to the dataset formed after parameter transformation, optimization, and region labeling of the input dataset through the transfer learning model. This dataset includes a subset of acoustic and optical parameters with region adaptation labels, as well as associated material type labels and temperature labels.

[0164] Step 503: Divide the crack at the junction of the base layer and the surface layer into a crack core area, a crack transition area, and a crack periphery area. Extract the converted parameters corresponding to each area from the parameter conversion dataset. Combine the reference correspondence to calculate the preliminary width value of multiple sampling points in each area.

[0165] In this step, the core area of ​​the crack refers to the main part of the crack at the junction of the base layer and the surface layer, which is the widest and most complete in shape.

[0166] The crack transition zone refers to the middle part connecting the core area of ​​the crack and the surrounding area of ​​the crack. Its width is between that of the core area and the surrounding area, and it is gradually changing.

[0167] The area surrounding a crack refers to the outer edge of the crack, where the width gradually decreases until it disappears; it is the extended area of ​​the crack.

[0168] The converted parameters refer to the standardized parameters extracted from the parameter conversion dataset that are adapted to each crack region. They are acoustic and optical parameters that have undergone numerical adjustment, unit conversion, and optimization.

[0169] The preliminary width value refers to the estimated crack width calculated for each sampling point based on the correlation between parameters and crack width in the benchmark correspondence. It has not undergone parameter correction for the temperature adaptation range.

[0170] In this embodiment, image segmentation and spatial partitioning methods can be used to divide the cracks at the junction of the base layer and the surface layer into a crack core region, a crack transition region, and a crack periphery region based on their location and morphological differences. Then, the transformed parameters corresponding to each region are extracted from the parameter transformation dataset. Combining this with the correlation between the parameters and crack width in the baseline correspondence, multiple sampling points are uniformly selected within each region. Abnormal sampling points exceeding the reasonable range for crack width corresponding to the pavement material are removed. For each retained valid sampling point, a preliminary width value is calculated. The aforementioned reasonable crack width range is determined based on a preset standard crack width interval in the baseline correspondence.

[0171] The initial width value can be obtained by multiplying the converted acoustic feature parameters by the first correlation coefficient in the baseline correspondence and the converted optical feature parameters by the second correlation coefficient in the baseline correspondence. Here, the converted acoustic feature parameters refer to the acoustic indices in the parameter-converted dataset that are adapted to the region to which the sampling point belongs, such as the standardized peak signal amplitude and signal propagation time; the converted optical feature parameters refer to the optical indices in the parameter-converted dataset that are adapted to the region to which the sampling point belongs, such as the standardized crack edge spacing.

[0172] The first correlation coefficient is a fixed coefficient preset in the benchmark correspondence that characterizes the linear correlation between acoustic parameters and crack width. The second correlation coefficient is a fixed coefficient preset in the benchmark correspondence that characterizes the linear correlation between optical parameters and crack width. Both correlation coefficients are determined based on the calibration data of preset standard crack samples.

[0173] Step 504: Adjust the initial width value of each sampling point according to the parameter correction coefficients corresponding to different temperature adaptation intervals in the reference correspondence, obtain the adjusted width value, and calculate the difference between the adjusted width values ​​of adjacent sampling points in each region.

[0174] In this embodiment, parameter correction coefficients corresponding to different temperature adaptation ranges can be extracted from the benchmark correspondence. The parameter correction coefficients are obtained by dividing the parameter calibration value at the standard temperature by the measured parameter value at the current temperature adaptation range. Here, the standard temperature refers to the preset pavement detection benchmark temperature in the benchmark correspondence, the parameter calibration value is the preset acoustic and optical parameter benchmark value of a standard crack sample at the standard temperature, and the measured parameter value is the measured acoustic and optical parameter value of the preset standard crack sample at the current temperature adaptation range.

[0175] Subsequently, based on the temperature labels of the input dataset, the current temperature adaptation range and corresponding correction coefficient are determined. The correction coefficient needs further calibration by considering the rate of change of the material's elastic properties and the rate of change of its optical refractive index within the current temperature adaptation range. Calibration is performed by multiplying the initial parameter correction coefficient, the material's elastic property stability coefficient, and the material's optical refractive index stability coefficient. The material's elastic property stability coefficient and the material's optical refractive index stability coefficient are both preset coefficients in the benchmark correspondence that match the temperature adaptation range. Then, the initial width value of each sampling point is multiplied by this correction coefficient to obtain the adjusted width value. Simultaneously, the difference in adjusted width values ​​between adjacent sampling points within each region is calculated.

[0176] Step 505: Based on the adjusted width difference between adjacent sampling points, smooth the adjusted width values ​​of the crack core area and crack transition area, and the adjusted width values ​​of the crack transition area and crack periphery area to form segmented width data for each area.

[0177] In this step, the segmented width data refers to the continuous data formed after smoothing the adjusted width values ​​of the crack core area, transition area, and surrounding area. This data presents the crack width distribution by region.

[0178] In this embodiment, the smoothness of the width change between the core area of ​​the crack and the transition area of ​​the crack, and between the transition area of ​​the crack and the surrounding area of ​​the crack, is first determined based on the adjusted width difference between adjacent sampling points. Then, the connection points with excessively large adjusted width differences are smoothed to ensure that the adjusted width values ​​of different areas are continuously connected, avoiding abrupt changes, and finally forming continuous segmented width data for each area.

[0179] Step 506: Based on the segment width data of each region, calculate the actual crack width at the junction of the base layer and the surface layer.

[0180] In this embodiment of the application, the average value of the adjusted width of all valid sampling points is calculated based on the segmented width data of each region. The average value is then corrected by combining the area ratio of each region, and the actual crack width at the junction of the base layer and the surface layer is finally obtained.

[0181] Optionally, step 502 involves using a transfer learning model to transform the input dataset into a parameter transformation dataset, specifically including the following steps:

[0182] Step 511: Extract parameter conversion rules from the benchmark correspondence that match the material type label and temperature label of the input dataset.

[0183] In this embodiment of the application, it is necessary to select parameter conversion rules that completely match the material type labels and temperature labels of the input dataset from the benchmark correspondence. These rules are based on historical detection data of the same material type and temperature range. The rules include the numerical conversion ratio of acoustic feature parameters, the unit conversion standard of optical parameters, and parameter adaptation thresholds.

[0184] Step 512: Extract the target parameter processing logic that matches the material type label and temperature label from the pre-stored pavement crack detection transfer framework in the transfer learning model.

[0185] In this step, the pavement crack detection transfer framework refers to the parameter processing framework pre-stored in the transfer learning model, designed for different pavement crack detection scenarios. This transfer framework includes target parameter processing logic corresponding to various material types and temperature ranges.

[0186] The target parameter processing logic refers to the set of parameter processing rules extracted from the pavement crack detection migration framework that matches the current material type and temperature range. This target parameter processing logic specifies the order and method of parameter splitting, adjustment, and optimization.

[0187] In this embodiment, the target parameter processing logic corresponding to the material type label and temperature label of the input dataset is extracted from the pavement crack detection transfer framework pre-stored in the transfer learning model. This logic specifies the processing order, adjustment method and optimization criteria of acoustic parameters and optical parameters to ensure that the parameter processing process is adapted to the material properties and temperature conditions.

[0188] Step 513: Based on the target parameter processing logic and the parameter conversion rules, perform numerical adaptation adjustment on the first acoustic feature parameter subset in the input dataset to generate the second acoustic feature parameter subset, and perform unit conversion on the first optical parameter subset in the input dataset to generate the second optical parameter subset.

[0189] In this step, the first subset of acoustic feature parameters refers to the set of raw acoustic feature parameters split from the input dataset. This subset includes unprocessed acoustic metrics such as signal amplitude peak and signal propagation time.

[0190] The second subset of acoustic feature parameters refers to the set of acoustic parameters obtained by numerically adapting and adjusting the first subset of acoustic feature parameters according to the parameter transformation rules, thus eliminating the numerical scale differences of the acoustic parameters.

[0191] The first subset of optical parameters refers to the original set of optical parameters separated from the input dataset. This subset includes unprocessed optical parameters such as crack edge spacing and is the initial optical data for parameter transformation.

[0192] The second subset of optical parameters refers to the set of optical parameters obtained by converting the first subset of optical parameters to units according to the parameter conversion rules, which realizes the unification of optical parameter units.

[0193] In this embodiment, the input dataset is first split into a first subset of acoustic feature parameters and a first subset of optical parameters. Based on the target parameter processing logic and parameter conversion rules, the first subset of acoustic feature parameters is adapted and adjusted according to a preset numerical conversion ratio to generate a second subset of acoustic feature parameters. Then, the first subset of optical parameters is converted into a second subset of optical parameters using a unit conversion standard with unified physical units.

[0194] Step 514: Based on the historical conversion experience information of detection parameters, optimize the second acoustic feature parameter subset and the second optical parameter subset through a transfer learning model to obtain the third acoustic feature parameter subset and the third optical parameter subset.

[0195] In this step, the third acoustic feature parameter subset refers to the acoustic parameter set obtained by combining the historical detection parameter transformation experience information with the transfer learning model to perform error correction and accuracy optimization on the second acoustic feature parameter subset.

[0196] The third optical parameter subset refers to the set of optical parameters obtained by combining historical detection parameter conversion experience information with a transfer learning model to correct errors and optimize accuracy of the second optical parameter subset.

[0197] In this embodiment, historical conversion experience information of detection parameters is first retrieved. This conversion experience information includes optimized cases of parameter processing under the same materials and temperature range, error correction data, and successful experiences. Then, based on this conversion experience information, the parameter optimization module of the transfer learning model performs error correction and accuracy improvement processing on the second acoustic feature parameter subset and the second optical parameter subset to obtain a third acoustic feature parameter subset and a third optical parameter subset with higher accuracy.

[0198] Step 515: Based on the parameter adaptation threshold in the parameter conversion rules, assign a region adaptation identifier to each parameter in the third acoustic feature parameter subset and the third optical parameter subset to form the fourth acoustic feature parameter subset and the fourth optical parameter subset;

[0199] In this step, the parameter adaptation threshold refers to the parameter value limit set in the parameter conversion rules for dividing the crack region. This threshold is set based on the parameter characteristics of different regions.

[0200] The region adaptation identifier refers to the identifier assigned to each parameter in the third acoustic feature parameter subset and the third optical parameter subset, representing the crack region to which it belongs.

[0201] The fourth subset of acoustic feature parameters refers to the third subset of acoustic feature parameters with region adaptation identifiers, which clearly defines the crack region corresponding to each acoustic parameter.

[0202] The fourth optical parameter subset refers to the third optical parameter subset with region adaptation identifiers, which clearly defines the crack region corresponding to each optical parameter.

[0203] In this embodiment, the crack region corresponding to each parameter in the third acoustic feature parameter subset and the third optical parameter subset can be determined according to the parameter adaptation threshold in the parameter conversion rule, and a corresponding region adaptation identifier can be assigned to each parameter to form a fourth acoustic feature parameter subset and a fourth optical parameter subset with region adaptation identifiers.

[0204] Step 516: Associate and integrate the fourth acoustic feature parameter subset and the fourth optical parameter subset with the corresponding material type label and temperature label to form a parameter conversion dataset.

[0205] In this embodiment, the fourth acoustic feature parameter subset and the fourth optical parameter subset are first associated and bound with the corresponding material type label and temperature label to obtain associated data, ensuring that each parameter can be traced back to its material and temperature background. Then, all associated data are categorized and organized according to regional adaptation identifiers, and integrated to form a parameter conversion dataset with a clear structure and strong adaptability.

[0206] This application's embodiments address the issues of low distributed detection accuracy and poor regional connectivity in existing solutions by using regional processing and temperature and material adaptation corrections, thus providing accurate and consistent results for the detection of hidden crack widths.

[0207] Figure 3 This is a schematic diagram of a specific implementation of a distributed detection system for surface crack width in linear road engineering provided in this application, referring to... Figure 3 The system may include:

[0208] Acquisition module 21 is used to acquire ambient temperature data, acoustic signals of cracks at the junction of the base layer and the surface layer, and optical images of cracks in the road to be inspected.

[0209] Processing module 22 is used to perform noise reduction processing on the acoustic signal and the crack optical image to obtain an effective acoustic signal and an effective optical image;

[0210] Correction module 23 is used to correct the propagation attenuation deviation of the effective acoustic signal and the imaging distortion deviation of the effective optical image based on the ambient temperature data, the material mechanical parameters and optical property parameters of the pavement material corresponding to the road to be tested, and the size calibration relationship of the crack optical image, using multi-physics coupling technology, to obtain a calibrated acoustic signal and a calibrated optical image.

[0211] The association module 24 is used to associate and fit the acoustic feature parameters of the calibration acoustic signal and the crack edge spacing of the calibration optical image with the sample crack width of the preset standard crack sample to establish a benchmark correspondence.

[0212] The inversion module 25 is used to perform distributed width inversion on the crack at the junction of the base layer and the surface layer based on the acoustic feature parameters, the crack edge spacing and the reference correspondence, and obtain the actual crack width.

[0213] This application provides a distributed detection system for the width of surface cracks in linear road engineering projects, which is used to implement the aforementioned distributed detection method for the width of surface cracks in linear road engineering projects. Therefore, the specific implementation of the distributed detection system for the width of surface cracks in linear road engineering projects can be found in the embodiment section of the distributed detection method for the width of surface cracks in linear road engineering projects described above. The specific implementation can be referred to the description of the corresponding embodiments, and will not be repeated here.

[0214] This application also provides an electronic device, comprising: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of the distributed detection method for crack width on the surface of a linear engineering road as described above.

[0215] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the distributed detection method for surface crack width of linear engineering road described above.

[0216] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, random access memory, portable hard drives, magnetic disks, or optical disks.

[0217] The embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the above embodiments of the distributed detection method for surface crack width in linear road engineering.

[0218] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. 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.

[0219] The above provides a detailed description of the distributed detection method, system, equipment, and medium for surface crack width in linear road engineering provided by this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A distributed detection method for the width of surface cracks in linear road engineering projects, characterized in that, include: Collect ambient temperature data, acoustic signals of cracks at the junction of the base and surface layers, and optical images of the cracks of the road to be inspected; The acoustic signal and the crack optical image are denoised to obtain an effective acoustic signal and an effective optical image; Based on the ambient temperature data, the material mechanical parameters and optical property parameters of the pavement material corresponding to the road to be tested, and the size calibration relationship of the crack optical image, the propagation attenuation deviation of the effective acoustic signal and the imaging distortion deviation of the effective optical image are corrected by multi-physics coupling technology to obtain the calibrated acoustic signal and the calibrated optical image. The acoustic feature parameters of the calibration acoustic signal and the crack edge spacing of the calibration optical image are correlated and fitted with the sample crack width of the preset standard crack sample to establish a benchmark correspondence. Based on the acoustic feature parameters, the crack edge spacing, and the baseline correspondence, a distributed width inversion of the crack at the junction of the base layer and the surface layer is performed using a transfer learning model to obtain the actual crack width. Based on the ambient temperature data, the material mechanical parameters and optical property parameters of the pavement material corresponding to the road under test, and the size calibration relationship of the crack optical image, multi-physics coupling technology is used to correct the propagation attenuation deviation of the effective acoustic signal and the imaging distortion deviation of the effective optical image, resulting in a calibrated acoustic signal and a calibrated optical image, including: Based on the ambient temperature data and the material mechanical parameters of the pavement material corresponding to the road to be tested, the first correlation between the propagation attenuation deviation of the effective acoustic signal and the material elastic properties and ambient temperature is determined using multi-physics coupling technology. Based on the first correlation, the propagation amplitude change and propagation time offset in the effective acoustic signal are adjusted to obtain the attenuation correction amount of the effective acoustic signal; Based on the material mechanical parameters, optical property parameters and ambient temperature data of the pavement material corresponding to the road to be detected, a second correlation is determined between the imaging distortion deviation of the effective optical image and the material optical refractive index, mechanical deformation and ambient temperature. Based on the imaging distortion deviation, the second correlation relationship, and the size calibration relationship of the slit optical image, the pixel stretching and edge offset in the effective optical image are adjusted to obtain the distortion correction amount of the effective optical image; The effective acoustic signal and the effective optical image are corrected according to the attenuation correction amount and the distortion correction amount, respectively, to obtain the calibrated acoustic signal and the calibrated optical image; Based on the acoustic feature parameters, the crack edge spacing, and the baseline correspondence, a distributed width inversion of the crack at the junction of the base layer and the surface layer is performed using a transfer learning model to obtain the actual crack width, including: Based on the temperature adaptation range and material adaptation parameters in the benchmark correspondence, the acoustic feature parameters and the crack edge spacing are weighted and integrated to form an input dataset, and the corresponding material type label and temperature label are marked in the input dataset; The input dataset is transformed into a parameter-transformed dataset using a transfer learning model. The crack at the junction of the base layer and the surface layer is divided into a crack core area, a crack transition area, and a crack periphery area. The converted parameters corresponding to each area are extracted from the parameter conversion dataset. Based on the reference correspondence, the preliminary width values ​​of multiple sampling points in each area are calculated. Based on the parameter correction coefficients corresponding to different temperature adaptation ranges in the aforementioned benchmark correspondence, the initial width value of each sampling point is adjusted to obtain the adjusted width value, and the difference between the adjusted width values ​​of adjacent sampling points in each region is calculated. Based on the adjusted width difference between adjacent sampling points, the adjusted width values ​​of the crack core area and crack transition area, and the adjusted width values ​​of the crack transition area and crack periphery area are smoothed to form segmented width data for each area. Based on the segmented width data of each region, the actual crack width at the junction of the base layer and the surface layer is calculated.

2. The method according to claim 1, characterized in that, Based on the ambient temperature data and the material mechanical parameters of the pavement material corresponding to the road to be tested, a first correlation between the propagation attenuation deviation of the effective acoustic signal and the material elastic properties and ambient temperature is determined using multiphysics coupling technology, including: Elastic correlation information is extracted from the material mechanics parameters of the pavement material corresponding to the road to be tested. The elastic correlation information includes the elastic response value and deformation recovery rate of the pavement material under different stresses. Based on the ambient temperature data, multiple temperature gradient intervals are divided, and the fluctuation range of elastic response values ​​and the trend of deformation recovery rate within each temperature gradient interval are analyzed to generate target elastic information. Attenuation feature data is extracted from the effective acoustic signal, and the attenuation feature data includes the amplitude reduction during signal propagation and the duration extension during signal propagation. Using multi-physics coupling technology, the fluctuation amplitude and deformation recovery rate of the elastic response value corresponding to each temperature gradient interval in the target elastic information are matched with the amplitude reduction and duration extension. Combined with the influence of the thermal conductivity information of the pavement material of the road to be tested on temperature transmission, an elastic attenuation correlation table is formed. Based on the elastic attenuation correlation table, a first correlation is determined between the propagation attenuation deviation of the effective acoustic signal and the material elastic properties and ambient temperature.

3. The method according to claim 1, characterized in that, Based on the material mechanical parameters, optical property parameters, and ambient temperature data of the pavement material corresponding to the road to be tested, a second correlation is determined between the imaging distortion deviation of the effective optical image and the material's optical refractive index, mechanical deformation, and ambient temperature, including: Transmission correlation information is extracted from the optical property parameters of the pavement material of the road to be tested, and shape change information is extracted from the material mechanical parameters of the pavement material of the road to be tested. Ambient light intensity data is acquired when optical images of cracks are captured on the road to be detected. Combined with ambient temperature data, the magnitude of change in the light transmission correlation information and the rate of change in the shape change information are analyzed to generate change record information. The effective optical image is divided into a crack core region image and a crack periphery region image. The pixel stretching degree of the crack core region image and the edge offset distance of the crack periphery region image are detected under different ambient temperatures and ambient light intensities to generate regional distortion information. The amplitude of light transmission change and the rate of shape change in the change record information are matched with the degree of pixel stretching and the edge offset distance in the regional distortion information. Combined with the influence of the thermal conductivity information of the pavement material corresponding to the road under test on temperature transmission and the superposition effect of ambient light intensity on light refraction, a distortion association table is formed. Based on the distortion correlation table, a second correlation is determined between the imaging distortion deviation of the effective optical image and the material's optical refractive index, mechanical deformation, and ambient temperature.

4. The method according to claim 1, characterized in that, The input dataset is transformed into a parameter-transformed dataset using a transfer learning model, including: Extract parameter transformation rules that match the material type labels and temperature labels of the input dataset from the benchmark correspondence; Extract target parameter processing logic that matches the material type label and temperature label from the pre-stored pavement crack detection transfer framework in the transfer learning model; Based on the target parameter processing logic and the parameter conversion rules, the first acoustic feature parameter subset in the input dataset is numerically adapted and adjusted to generate the second acoustic feature parameter subset, and the first optical parameter subset in the input dataset is unit converted to generate the second optical parameter subset. Based on historical detection parameter conversion experience information, the second acoustic feature parameter subset and the second optical parameter subset are optimized through a transfer learning model to obtain the third acoustic feature parameter subset and the third optical parameter subset; According to the parameter adaptation threshold in the parameter conversion rule, a region adaptation identifier is assigned to each parameter in the third acoustic feature parameter subset and the third optical parameter subset to form the fourth acoustic feature parameter subset and the fourth optical parameter subset; The fourth acoustic feature parameter subset and the fourth optical parameter subset are associated and integrated with the corresponding material type labels and temperature labels to form a parameter transformation dataset.

5. The method according to claim 1, characterized in that, The acoustic characteristic parameters of the calibrated acoustic signal and the crack edge spacing of the calibrated optical image are correlated and fitted with the sample crack width of a preset standard crack sample to establish a benchmark correspondence, including: Based on the ambient temperature data, a simulated testing environment for the road to be tested is constructed, and standard acoustic signals and standard optical images of pre-set standard crack samples are collected in the simulated testing environment. The standard acoustic signal and the standard crack optical image are denoised to obtain the standard effective acoustic signal and the standard effective optical image; Based on the ambient temperature data, the material mechanical parameters and optical property parameters of the preset standard crack sample, and the size calibration relationship of the standard crack optical image, the propagation attenuation deviation of the standard effective acoustic signal and the imaging distortion deviation of the standard effective optical image are corrected to obtain the standard calibrated acoustic signal and the standard calibrated optical image. Standard acoustic feature parameters are extracted from the standard calibration acoustic signal. The standard acoustic feature parameters include the peak value of the signal amplitude and the signal propagation time. The pixel spacing between the two edges of the crack in the standard calibration optical image is used as the standard crack edge spacing. Based on the type of pavement material and the ambient temperature data, the correlation between the changing trends of the standard acoustic characteristic parameters, the standard crack edge spacing and the sample crack width of the preset standard crack sample is analyzed to establish a benchmark correspondence.

6. A distributed detection system for the width of cracks on the surface of linear road engineering projects, used to execute the distributed detection method for the width of cracks on the surface of linear road engineering projects as described in claim 1, characterized in that, include: The acquisition module is used to acquire ambient temperature data, acoustic signals of cracks at the junction of the base layer and the surface layer, and optical images of cracks in the road to be inspected. The processing module is used to perform noise reduction processing on the acoustic signal and the crack optical image to obtain an effective acoustic signal and an effective optical image; The correction module is used to correct the propagation attenuation deviation of the effective acoustic signal and the imaging distortion deviation of the effective optical image based on the ambient temperature data, the material mechanical parameters and optical property parameters of the pavement material corresponding to the road to be tested, and the size calibration relationship of the crack optical image, using multi-physics coupling technology, to obtain a calibrated acoustic signal and a calibrated optical image. The correlation module is used to correlate and fit the acoustic feature parameters of the calibration acoustic signal and the crack edge spacing of the calibration optical image with the sample crack width of the preset standard crack sample to establish a benchmark correspondence. The inversion module is used to perform distributed width inversion of the crack at the junction of the base layer and the surface layer based on the acoustic feature parameters, the crack edge spacing and the reference correspondence, and obtain the actual crack width through a transfer learning model.

7. A computing device, characterized in that, It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement a distributed detection method for surface crack width of linear engineering road as described in any one of claims 1 to 5.

8. A computer storage medium, characterized in that, The system contains a computer program that, when executed by a computer, implements a distributed detection method for the width of surface cracks in linear road engineering as described in any one of claims 1 to 5.

Images