Ground penetrating radar image quality evaluation method and system for asphalt pavement cavity disease identification

Abstract

The application discloses a ground penetrating radar image quality evaluation method and system for asphalt pavement cavity disease identification, and belongs to the technical field of ground penetrating radar images. In order to solve the problem that there is no quality evaluation method for ground penetrating radar images of asphalt pavement cavity diseases at present, the application is used for detecting a road, obtaining a detection B-scan image of the road, and establishing a corresponding control group forward model according to a design file of the road. The control group is obtained by changing the multi-layer asphalt pavement structure in the design file of the road into a single-layer pavement structure. The image simulated by the control group forward model is recorded as a control group B-scan image. Then, the image quality of the detection B-scan image is evaluated according to an SSIM-S index. The application is used for quality evaluation of ground penetrating radar images.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of ground penetrating radar images, and relates to a method and system for evaluating the quality of ground penetrating radar images of asphalt pavement cavity diseases. BACKGROUND

[0002] Ground penetrating radar (GPR) is a non-destructive and rapid detection technology that uses electromagnetic wave reflection to determine the distribution of underground media, and has become one of the preferred methods for detecting hidden diseases of roads. The key to accurately detecting hidden diseases of roads using GPR is GPR image recognition, but the quality of current ground penetrating radar images is uneven. The quality of ground penetrating radar images seriously affects the identification of hidden diseases of roads. The higher the quality of ground penetrating radar images, the more obvious the characteristics of hidden diseases in the images, and the higher the detection accuracy. Otherwise, it will lead to a large probability of misjudgment and omission, and cannot fully develop the potential of GPR technology. Therefore, the evaluation of the quality of ground penetrating radar images is a bottleneck that restricts the development of smart cities supported by the technology.

[0003] Currently, there is no unified standard and method for evaluating the quality of ground penetrating radar images. Therefore, when the image quality is poor, if there is no way to evaluate its quality, but all ground penetrating radar images are directly regarded as unified quality images without difference for identifying road cavity diseases, the identification effect of some road diseases will not be ideal. SUMMARY

[0004] The present application is to solve the problem that there is no method for evaluating the quality of ground penetrating radar images of asphalt pavement cavity diseases.

[0005] A method for evaluating the quality of ground penetrating radar images for identifying asphalt pavement cavity diseases, comprising the following steps:

[0006] For detecting a road, a ground penetrating radar image of the detected road, i.e., a detected B-scan image, is obtained, and a corresponding control group forward model is established according to a design file of the detected road. The control group is to change the multi-layer asphalt pavement structure in the design file of the detected road to a single-layer pavement structure. The image simulated by the control group forward model is recorded as a control group B-scan image. Then, the image quality of the detected B-scan image is evaluated according to the SSIM-S index.

[0007]

[0008] In the formula, denotes the structural similarity index of the image and the image , denotes the structural similarity index of the image and the image Spearman rank correlation coefficient index of images Spearman rank correlation coefficient index of images respectively are the detection B-scan image and the control group B-scan image.

[0009] Further, the forward model adopts a 400MHz Ricker wavelet as an excitation source signal.

[0010] Further, in the process of obtaining the ground penetrating radar image of the detection road, the antenna transmitting-receiving interval is set to be less than 25cm. Preferably, the antenna transmitting-receiving interval is set to be [10cm, 20cm].

[0011] Further, the antenna transmitting-receiving interval in the forward model is equal to the antenna transmitting-receiving interval set in the process of obtaining the ground penetrating radar image of the detection road.

[0012] A ground penetrating radar image quality evaluation system for asphalt pavement cavity disease identification, comprising:

[0013] A ground penetrating radar image acquisition unit: for a detection road, obtaining a ground penetrating radar image of the detection road, i.e. a detection B-scan image;

[0014] A control group forward model establishment unit: establishing a corresponding control group forward model according to the design document of the detection road, and the control group is to change the multi-layer asphalt pavement structure in the design document of the detection road to a single-layer pavement structure;

[0015] A control group image acquisition unit: based on the established control group forward model, a control group B-scan image is simulated and obtained;

[0016] A quality evaluation unit: performing image quality evaluation of the detection B-scan image according to the SSIM-S index;

[0017]

[0018] In the formula, indicates the structural similarity index of images and images indicates the structural similarity index of images and images Spearman rank correlation coefficient index of images , images respectively are the detection B-scan image and the control group B-scan image.

[0019] Further, the forward model adopts a 400MHz Ricker wavelet as an excitation source signal.

[0020] ​Further, in the process of acquiring the ground penetrating radar image for detecting the road, the antenna transmitting-receiving interval is set to be less than 25 cm.

[0021] Further, the antenna transmitting-receiving interval in the forward model is equal to the antenna transmitting-receiving interval set in the process of acquiring the ground penetrating radar image for detecting the road. Beneficial effects

[0022] The present application comprehensively evaluates the asphalt pavement cavity disease characteristics by calculating the SSIM of the B-scan images of the experimental group and the control group, and simultaneously using the SROCC between the waveform graphs of the experimental group and the control group as an auxiliary. The present application uses SSIM-S for evaluation, and the range of SSIM-S is [0, 1]. The larger the numerical value is, the higher the similarity of the image is, and the better the result of the forward simulation of the experimental group is. The present application not only considers the particularity of the B-scan image, but also solves the problem that when the target body and the background dielectric constant in the radar propagation process differ greatly, the medium has strong absorption of electromagnetic wave energy, and the B-scan image features are not obvious, which leads to inaccurate evaluation of the B-scan image quality of the experimental group. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 Blackman-Harris Window time domain waveform graph (400MHz and 800MHz);

[0024] Figure 2 Blackman-Harris Window spectrum graph (400MHz and 800MHz);

[0025] Figure 3 Ricker time domain waveform graph (400MHz and 800MHz);

[0026] Figure 4 Ricker spectrum graph (400MHz and 800MHz);

[0027] Figure 5 Parameter distribution diagram of UPML boundary condition;

[0028] Figure 6 Asphalt pavement cavity disease forward model diagram with depth direction change;

[0029] Figure 7 Asphalt pavement cavity disease forward model diagram corresponding to the change of transmitting-receiving interval;

[0030] Figure 8 Asphalt pavement cavity disease forward model diagram corresponding to the change of antenna transmitting-receiving source interval and cavity depth;

[0031] Figure 9 The forward model schematic diagram for the control group;

[0032] Figure 10 The B-scan data schematic diagram for the experimental group and the control group;

[0033] Figure 11 The display diagram for the B-scan feature not obvious;

[0034] Figure 12 The display diagram for the wave field snapshot with reflection signal;

[0035] Figure 13 The waveform diagram for the experimental group and the control group;

[0036] Figure 14 The flow chart for the SSIM-S index construction process and evaluation;

[0037] Figure 15 The three-dimensional fold surface diagram according to the data in Table 3;

[0038] Figure 16 The B-scan diagram for the distance of the cavity disease from the depth of the road base layer being 0 cm;

[0039] Figure 17 The SSIM-S curve diagram for the distance of the cavity disease from the depth of the road base layer;

[0040] Figure 18 The SSIM-S curve diagram for the distance of the cavity disease from the depth of the road base layer;

[0041] Figure 19 The wave field snapshot for different cavity disease depths when the distance between the antenna and the receiver is 5 cm;

[0042] Figure 20 The wave field snapshot for different antenna and receiver distances when the cavity disease depth is 25 cm. DETAILED DESCRIPTION

[0043] In order to solve the problems in the background art, the present application considers the ground penetrating radar B-scan data, combines the waveform diagram, and comprehensively calculates the image quality evaluation method in the field of computer vision to propose an index suitable for ground penetrating radar image quality evaluation. The present application will be described in detail below in combination with specific embodiments. DETAILED DESCRIPTION

[0045] The present embodiment is a ground penetrating radar image quality evaluation method for identifying cavity diseases of asphalt pavement.

[0046] Maxwell's equations are the theoretical foundation for quantitatively analyzing the detection performance of ground-penetrating radar (GPR). They describe the relationship between electric and magnetic fields and elucidate the propagation laws of high-frequency electromagnetic waves in a medium during GPR operation. FDTD (Flat-Difference Ground-Penetrating Radar) is used as a numerical simulation method for electromagnetic fields, decomposing the target into Yee elements. This form is similar to the resolution units of a computer, progressively approximating the final target. However, when solving electromagnetic problems, Maxwell's equations are insufficient to explain the magnitude of each electromagnetic field. Therefore, it is also necessary to consider the relationships between various field quantities and, in conjunction with constitutive relations, decompose the current density into conductor current density and applied current density.

[0047] The resulting system of equations consists of two vector equations. Each of these vector equations in three-dimensional space can be decomposed into three scalar equations, equivalent to six scalar equations. The equations are in the following form:

[0048]

[0049] In the formula, Represents electric field strength (V / m). This represents the magnetic field strength (A / m). This represents the free current density (A / m²). This represents the magnetic flux density (V / m²). , The values ​​represent the permeability and permittivity of the medium, with subscripts x, y, and z representing the components of the corresponding physical quantities in three-dimensional space.

[0050] In the Cartesian coordinate system, Maxwell's equations, after sampling their spatial components in the two-dimensional finite difference time-domain method, become two-dimensional finite difference equations:

[0051]

[0052]

[0053] in, Indicates at time step and spatial grid points The magnetic field component at that location; The time step; This is the spatial step size (in the y-direction); This represents the time step n and the spatial grid point. The electric field component at that location.

[0054] In FDTD numerical calculations, stability conditions affect the calculation accuracy. Under two-dimensional FDTD conditions, the stability condition for electromagnetic wave fluctuations in ground-penetrating radar is given by equation (10):

[0055]

[0056] In the formula, The speed of light, And The step length of two-dimensional Yee unit in X and Y direction, The time step of wave.

[0057] Excitation source and absorption boundary condition:

[0058] In forward simulation, excitation source is an important parameter. It will affect the electromagnetic wave behavior in a particular medium. Ground penetrating radar as a device for detecting hidden diseases of road surface, the main excitation source function is pulse function. Blackman-Harris Window and Ricker are one of the commonly used excitation source functions in FDTD method, and are widely used in simulation of seismic exploration and geological detection. The invention selects the above two excitation source functions for comparison and drawing. On this basis, at the same time, the time domain waveforms of 400MHz and 800MHz frequencies are compared.

[0059] The function expression of Blackman-Harris Window is as follows:

[0060]

[0061] In the formula, The center frequency of excitation source is a fixed parameter. Indicates time, The duration of excitation source function. T value is calculated by formula (19):

[0062]

[0063] In the formula, The center frequency of excitation source function.

[0064] Through programming calculation, the time domain waveform diagram and frequency spectrum diagram of the main frequency of 400Mhz and 800Mhz are obtained, as shown in Figures 1-2 .

[0065] Ricker wavelet is a zero-phase wavelet in radar signal simulation, and the expression is as follows:

[0066]

[0067] In the formula, The center frequency of excitation source function. e is natural logarithm. Pi is the ratio of circumference to diameter.

[0068] Based on the above, through programmed calculation, the time-domain waveform and spectrum diagram of the main frequency of 400Mhz and 800Mhz under the Ricker excitation source are obtained, as shown in Figures 3-4

[0069] Analysis Figures 1-2 and Figures 3-4 The time-domain waveform and spectrum of the excitation source function are analyzed. The Ricker wavelet is a waveform with a sharp pulse shape, which is similar to a bell-shaped curve. This shape makes the Ricker wavelet perform well in target detection and imaging. Because it can produce high-resolution echoes and provide better target positioning capability. There are two different side lobes in the waveform, and the center frequency of the signal source is more suitable for simulating complex propagation medium models, such as asphalt road structure layers. On the contrary, the Blackman-Harris pulse waveform is smoother and is not suitable for detecting sudden target bodies such as road cavity diseases. The Ricker wavelet has a wider frequency spectrum width than the Blackman-Harris pulse, which means it can provide better frequency resolution. Obviously, the Ricker wavelet is more representative when simulating the detection of asphalt road cavity diseases.

[0070] By comparing the time-domain waveforms and spectra of 400MHz and 800MHz, it is found that whether the excitation source is a Ricker wavelet or a Blackman-Harris pulse, 400MHz is more suitable for detecting hidden diseases in the road surface, especially when the disease depth is below the base layer. Figures 1-2 In the Blackman-Harris excitation source, the oscillation speed of 800MHz is twice that of 400MHz. This will result in faster propagation speed in the same medium, but the energy loss also increases, and the detection depth is greatly attenuated. Similarly, Figures 3-4 In the Ricker excitation source, the time-domain waveform of 800MHz is also narrower than that of 400MHz, which means the period is shorter and the oscillation speed is faster. From the perspective of the spectrum diagram, the frequency spectrum bandwidth of 800MHz is much larger than that of 400MHz. The wider the frequency spectrum bandwidth, the shorter the duration. Therefore, considering the location of the cavity disease in the asphalt pavement and the multi-layer structure of the asphalt pavement, the Ricker wavelet excitation source of 400MHz is used as the forward pulse model.

[0071] ​In the process of ground penetrating radar (GPR) based asphalt pavement detection, it is impossible to guarantee that the target body to be detected is regular in shape or has a broken boundary. Such problems may cause the spatial waveform to be truncated when simulating electromagnetic field by using FDTD, resulting in abnormal electromagnetic wave reflection. In order to make the forward result more accurate, a method of setting a truncated boundary grid for the model is provided to simulate the propagation effect of an infinite space in a limited grid space. The present application adopts a uniaxial anisotropic ideal matching layer (UPML) absorbing boundary condition. The method is suitable for uniaxial anisotropic media and performs wave field absorption at the boundary grid. The UPML calculation is time-consuming and the calculation formula is based on Maxwell's equation as follows:

[0072]

[0073] In the formula, , and respectively represent the uniaxial anisotropic parameter tensor. imaginary unit; angular frequency; dielectric constant.

[0074] The UPML of two-dimensional GPR waves has four plane regions and four edge regions, and only , , components, at this time, the expressions of three parameters of the UPML are as formula (16)-(18). The parameter distribution diagram of the UPML boundary condition is as shown in Figure 5 .

[0075]

[0076] In the formula, the parameters and are used to absorb the evanescent waves reaching the UPML layer; the parameters and are the attenuation factors of the UPML region. In order to make the wave be fully absorbed and the simulation effect be better, the parameters and should have a proper spatial distribution in the UPML region.

[0077]

[0078] wherein, conductivity at position i, dielectric loss factor at position i, maximum value of the conductivity, maximum value of the dielectric loss factor; The length scale (PML medium thickness) at spatial location i is generally taken as , The spatial step size; m is an exponent used to adjust the conductivity and dielectric loss factor variation generally taken as ; The dielectric constant.

[0079] The subsequent need for forward model to verify the SSIM-S index, the forward model needs to be established process, the above content is the theoretical basis of the forward model.

[0080] The forward model of the cavity disease of the asphalt pavement:

[0081] The forward model here is a simulation method, first simulate the asphalt pavement structure and cavity disease through the computer, then get the simulated ground penetrating radar image through calculation, verify the SSIM-S index through the forward model. The new evaluation index is also inspired by the forward model, so the relationship between the forward model and the image quality evaluation index determines the SSIM-S index of the present application. At the same time, because the cavity disease characteristics of the experimental group in the forward model are affected by the multi-layer structure of the road, its characteristics are definitely not standard and obvious, the road structure in the control group model only has one layer structure, which greatly reduces other influencing factors, and the simulated cavity disease characteristics are relatively more standard and obvious. Therefore, by calculating the SSIM-S index of the two groups of models, the image quality of the experimental group model can be obtained.

[0082] The present application constructs three different forward models for detecting the cavity disease of the asphalt pavement by ground penetrating radar. As shown in Figures 6-8 The first layer is the surface layer, and the material is asphalt mixture. The second layer is the base layer, and the material is cement stabilized gravel. The third layer is the roadbed, and the main material is compacted soil and gravel. TA is the ground penetrating radar transmitting antenna, and RA is the radar receiving antenna. When the electromagnetic wave propagates between multiple different media, the distance between the antenna transmitting and receiving sources will affect the final signal processing result, and the burial depth of the target body will affect the detection accuracy. Therefore, the present application analyzes the influence of different depths of cavity disease and different antenna transmitting and receiving distances on the detection effect of the cavity disease of the asphalt pavement.

[0083] First, simulate the cavity disease of the asphalt pavement with different depths. The pavement structure is divided into the most common three-layer model. The Yee unit size is selected as 1 cm, and the z-direction size is ignored. The center frequency of the radar excitation source is set to 400Mhz, which is the best frequency for the application effect of the cavity disease of the asphalt pavement in actual detection. The cavity disease is Figures 6-8cube in the middle, with a boundary length of 0.3 m. Generally speaking, cavity disease refers to the cavity formed by the peeling of the material in the base and below the base. Therefore, in order to comprehensively simulate the position of cavity disease in the asphalt pavement, the present application is placed in three different layer positions. As shown in Figure 6 the first model, the distance x between the cavity disease and the top of the base is 0 m. Each time the depth is stepped down by 0.05 m, and so on, and finally the distance between the cavity disease and the top of the base in the final model is 0.4 m. The excitation source is Ricker, and the transmission source position starts at 0.1 m from the left end of the surface layer. After each transmission source transmits a signal, the transceiver source steps to the right with a step distance of 0.01 m. The boundary condition is UPML.

[0084] Secondly, forward modeling of different antenna transceiver spacings. As shown in Figure 7 the first model, the distance d between the antenna transceiver sources is 0.05 m. After the forward modeling is completed, the distance between the transceiver sources is changed, and the transceiver spacing d is stepped by 0.05 m, and finally d is 1 m.

[0085] Figure 8 the relationship between the changing antenna transceiver source spacing and the cavity depth.

[0086] The present application establishes a total of 180 asphalt road cavity disease forward modeling models, and the parameters such as the antenna spacing and the cavity depth of each model are shown in Table 1.

[0087] Table 1 Asphalt pavement cavity disease forward modeling model

[0088]

[0089] The electromagnetic parameters of each structural layer material of the asphalt pavement are shown in Table 2. The cavity disease is mainly composed of air, so the electromagnetic parameter values are consistent with air.

[0090] Table 2 Electromagnetic parameters of each structural layer and cavity of the asphalt pavement forward modeling model

[0091]

[0092] In order to compare the optimal antenna transceiver source distance of ground penetrating radar, the present application also simultaneously constructs 180 control group models. The control group models correspond one-to-one with the asphalt pavement cavity disease forward modeling models. The control group model schematic diagram is as shown in Figure 9The B-scan data of the experimental group and the control group are shown in FIG. 6.

[0093] Forward modeling results:

[0094] According to the established experimental group and the control group of the asphalt pavement cavity disease model, the radar scanning data is simulated. The data is A-scan. The present application integrates multiple A-scans by programming to obtain B-scan data of the experimental group and the control group of the asphalt pavement cavity disease, part of which is shown in FIG. 6. Figure 10 Figure 10 The left column of images is the experimental group image, and the right column is the control group image.

[0095] The B-scan data is a time-domain data form, which displays the echo signals in the underground medium as brightness or color changing with time. Generally, the X-axis represents the measurement position, the Y-axis represents the time, and the brightness or color represents the echo signal intensity or amplitude. The cavity disease of the experimental group model shows a non-standard hyperbolic feature on the B-scan, and the internal waveform oscillation is relatively chaotic. With the change of time domain, the diffraction characteristics on both sides of the cavity disease of the experimental group model do not tend to converge. The hyperbolic feature of the cavity disease of the control group model is standard, and the internal waveform oscillation is beautiful. The diffraction characteristics on both sides do not exceed the top amplitude.

[0096] The evaluation index of the forward data is as follows:

[0097] Generally, the B-scan data can reflect the signal characteristics of the hidden disease of the asphalt pavement and the surrounding environmental noise. The stronger the signal received by the radar receiving antenna, the more obvious the feature on the B-scan image. In the actual detection process, professionals also use the features on the B-scan image as the basis for on-site exploration. The number of pavement structure layers of the control group forward model is one, compared with the experimental group model, the loss of radar signal propagation in its interior will be greatly reduced. The B-scan data obtained by the control group model is also cleaner and more accurate, which can be used as a benchmark to measure the quality of B-scan data. The present application uses SSIM to calculate the correlation of B-scan data of the experimental group and the control group model.

[0098] ​SSIM is an index used to measure the similarity of two images. It is more consistent with the human eye's judgment of image quality than the traditional image quality measurement index. When the input image size is equal, the SSIM operation is shown in equation (23).

[0099]

[0100] In the formula, and are the average values of the image and the average value of the image y , and are the standard deviations of the two images, is the covariance between the two images, and are constants. The image represents the detection B-scan image (experimental group), and the image y represents the B-scan image obtained by simulating the forward model of the control group;

[0101] It should be noted that in the actual detection process, the ground penetrating radar actual image will be obtained through the foregoing content, and then according to the design file of the detected road, the corresponding forward model can be established. The average value of the ground penetrating radar actual B-scan image is ; the average value of the image obtained by simulating the corresponding forward model is ; thus, the SSIM index can be calculated; similarly, the subsequent SSIM-S index can also be calculated.

[0102] The SSIM range is [0, 1], and the larger the value, the higher the image similarity and the better the quality. However, for some special images, similar to medical CT and ground penetrating radar B-scan data, SSIM is not flexible enough and is not sensitive enough to image changes. At the same time, SSIM is not accurate enough for images with large gray scale differences. The gray scale contrast difference between strong amplitude and weak amplitude in the B-scan image is large. During the research, it was also found that a few forward model simulations observed that the cavity signal had been reflected to the receiving antenna in the wave field snapshot, but the cavity feature was not obvious on the B-scan image, as shown in Figures 11-12 , it can be seen that the B-scan feature is not obvious, Figure 11 , and the wave field snapshot has a reflected signal. Figure 12

[0103] ​Therefore, considering the above-mentioned problems, this invention introduces the Spearman rank correlation coefficient (SROCC) on the basis of SSIM to comprehensively evaluate the forward simulation results. SROCC is a non-parametric index that measures the correlation between two variables. It is also the Pearson correlation coefficient between rank variables. Its calculation formula is shown in equations (24) and (25).

[0104]

[0105] In the formula, The number of observed samples; Representing an image The i-th observation level, Indicates the average grade of the corresponding observations; Representing an image y The level of the i-th observation, This represents the average value of the corresponding observations; the observations are feature vectors of the image.

[0106] SROCC values ​​range from -1 to 1. Coefficients close to 1 and -1 indicate a high degree of similarity. This invention draws on the waveform diagrams of seismic waves during propagation to establish forward modeling waveform diagrams for asphalt pavement void defects based on forward modeling data. This invention generates all waveform diagrams for the experimental and control groups by reading forward modeling data and interpreting it through programming. A portion of these waveform diagrams are shown below. Figure 13 As shown, Figure 13 The left column of images represents the experimental group, and the right column represents the control group.

[0107] Due to the presence of gain and the influence of multiple layers in the experimental group model, the amplitude of the cavity defect waveform in the experimental group is stronger than that in the control group. However, this strong amplitude does not change the standard characteristics of the cavity defect. Therefore, this invention comprehensively evaluates the characteristics of asphalt pavement cavity defects by calculating the SSIM of the B-scan images of the experimental and control groups, and by using the SROCC between the waveforms of the experimental and control groups as an auxiliary factor. This invention proposes a new evaluation index, SSIM-S, based on the ground-penetrating radar image of asphalt pavement cavity defects. The SSIM-S calculation formula is shown in equation (26).

[0108]

[0109] The SSIM-S value ranges from [0, 1]. A higher value indicates a higher similarity between the images, resulting in better forward modeling results for the experimental group. SSIM-S not only considers the specific characteristics of B-scan images but also addresses the issue of inaccurate evaluation of the experimental group's B-scan image quality when the dielectric constants of the target and background differ significantly during radar propagation. This leads to stronger absorption of electromagnetic wave energy by the medium, resulting in less distinct simulated B-scan image features. The SSIM-S evaluation index is illustrated below.Figure 14 Firstly, the SSIM is calculated by the B-scan images of the experimental group and the control group forward model. Secondly, the Spearman rank correlation coefficient is calculated according to the waveform diagram of the two groups. Finally, the SSIM-S is obtained by substituting formula (26).

[0110] Based on the SSIM-S of the asphalt pavement cavity disease ground penetrating radar forward simulation evaluation:

[0111] The present application calculates the SSIM-S index according to the B-scan image and waveform diagram corresponding to each model through 360 forward models of the experimental group and the control group.

[0112] Table 3 SSIM-S index of each model

[0113]

[0114] In order to facilitate the display of data in Table 3, a three-dimensional folded surface diagram is prepared, as shown in Figure 15 . Figure 15 The SSIM-S, the distance between the antenna and the transceiver and the depth of the cavity disease from the pavement base are respectively taken as the Z axis, the X axis and the Y axis.

[0115] From Figure 15 It can be observed that the whole three-dimensional folded diagram presents a ridge-like shape. As the TA and RA distance increases and the cavity disease depth becomes shallower, the overall SSIM-S value is lower. This shows that the larger the antenna transceiver distance is, the less suitable it is for detecting the cavity disease close to the road surface. Such detection results will appear when the cavity disease feature and the asphalt pavement layer feature are mixed, and the diffraction wave feature at both ends of the cavity disease is almost disappeared. As shown in Figure 16 Therefore, in actual detection, the ground penetrating radar antenna related parameters can be better set according to this characteristic.

[0116] In the three-dimensional folded surface diagram, the present application projects the graph vertically on the YZ plane, and according to the interpolation method, the SSIM-S index curve of each antenna transceiver distance under different cavity disease depth ranges is obtained. As shown in Figure 17 .

[0117] From Figure 17It can be observed that in 20 different antenna-transceiver spacings, the SSIM-S value of the antenna-transceiver spacing of 5 cm is the highest, and the effect of the target image is the best. With the increase of the antenna-transceiver spacing, the SSIM-S value of the same depth of the cavity disease decreases, and the B-scan image quality becomes poor. When the antenna-transceiver spacing is in [5 cm, 25 cm], with the increase of the depth of the cavity disease, the SSIM-S value of the forward model increases. When the antenna-transceiver spacing is in [30 cm, 100 cm], with the increase of the depth of the cavity disease, the SSIM-S value of the forward model fluctuates repeatedly. The shorter the antenna-transceiver spacing, the longer the time consumption of the forward simulation, and the higher the accuracy. Therefore, considering the accuracy and time of the model simulation, the antenna-transceiver spacing of [10 cm, 20 cm] can be selected.

[0118] In the three-dimensional folded surface diagram, the graph is vertically projected on the XZ plane, and the SSIM-S index curve of the cavity disease at different depths under different antenna-transceiver spacings is obtained according to the interpolation method. Figure 18 As shown in the figure.

[0119] From Figure 18 It can be observed that in 9 different depths of the cavity disease, the SSIM-S value of the cavity disease 40 cm away from the road base is relatively the highest, and the characteristics of the target image are the best. With the increase of the depth of the cavity disease, the SSIM-S value of the same antenna-transceiver spacing becomes larger, which means that the B-scan image quality becomes better. When the antenna-transceiver spacing is less than 25 cm, the SSIM-S is positively correlated with the depth of the cavity disease and inversely correlated with the antenna-transceiver spacing. When the antenna-transceiver spacing is greater than 25 cm, the above rule changes. Therefore, when the antenna-transceiver spacing cannot be changed in actual detection or forward simulation, the deeper cavity disease is detected as much as possible. In this way, the detection accuracy can be improved to the greatest extent.

[0120] The above reasons are that the increase of the antenna-transceiver spacing first leads to the increase of the time of the direct wave reaching the antenna receiving end. The direct wave, the road surface reflected wave and the cavity disease interface reflected wave reach the antenna receiving end together. In the B-scan image, the characteristics of the three waveforms are mixed together, and it is difficult to distinguish the bit relationship of the asphalt pavement. Secondly, the deeper the depth of the cavity disease, the slower the propagation speed of the electromagnetic wave affected by the road surface medium, which leads to the longer time of the disease signal reaching the antenna receiving end. The arrival time of the disease signal is generally longer than that of the direct wave and the road surface bit reflected wave. The signal interference among the three is reduced. Figure 19 As shown in the figure.

[0121] Therefore, in Figure 18The results show that the deeper the hollow disease, the higher the SSIM-S of most antenna-transceiver spacings. When the antenna-transceiver spacing is in [60cm, 85cm], the arrival time of the disease signal and the direct wave signal is approximately the same. As shown in FIG. 8, when the antenna-transceiver spacing is 60cm and 85cm, the arrival time of multiple radar signals at the receiving end is approximately the same. Figure 20 The specific embodiments will be described below. DETAILED DESCRIPTION

[0123] The embodiment is a ground penetrating radar image quality evaluation system for hollow disease identification of asphalt pavement, which is a computer program corresponding to the ground penetrating radar image quality evaluation method for hollow disease identification of asphalt pavement described in embodiment one. More specifically, the ground penetrating radar image quality evaluation system for hollow disease identification of asphalt pavement described in the embodiment includes:

[0124] The ground penetrating radar image acquisition unit acquires the ground penetrating radar image of the detection road, i.e., the detection B-scan image. During the acquisition of the ground penetrating radar image of the detection road, the antenna-transceiver spacing is set to be less than 25cm, and preferably, the antenna-transceiver spacing is set to be [10cm, 20cm].

[0125] The control group forward model establishment unit establishes a corresponding control group forward model according to the design document of the detection road. The control group is to change the multi-layer asphalt pavement structure in the design document of the detection road to a single-layer pavement structure. The forward model uses a 400MHz Ricker wavelet as an excitation source signal. The antenna-transceiver spacing in the forward model is equal to the antenna-transceiver spacing set during the acquisition of the ground penetrating radar image of the detection road.

[0126] The control group image acquisition unit simulates the control group B-scan image based on the established control group forward model.

[0127] The quality evaluation unit evaluates the image quality of the detection B-scan image according to the SSIM-S index.

[0128]

[0129] In the formula, SSIM-S represents the structural similarity index of image and image. In the formula, S represents the Spearman rank correlation coefficient of image and image.

[0130] ​​​​​​​The present application can have other various embodiments, and those skilled in the art can make various corresponding changes and modifications according to the present application without departing from the spirit and essence of the present application, and these corresponding changes and modifications shall all belong to the protection scope of the claims of the present application.

Claims

1. A method for ground penetrating radar image quality assessment for asphalt pavement void disease identification, characterized in that, The method comprises the following steps: For the detection road, a ground penetrating radar image of the detection road, i.e., a detection B-scan image, is acquired, and a corresponding control group forward model is established according to a design file of the detection road, wherein the control group is a single-layer pavement structure obtained by changing a multi-layer asphalt pavement structure in the design file of the detection road; an image simulated by the control group forward model is denoted as a control group B-scan image; and then image quality evaluation of the detection B-scan image is performed according to an SSIM-S index; In the formula, representative image representative image representative image representative image representative image representative image representative image representative image The forward model of the asphalt pavement cavity disease: The forward model is a simulation method, which firstly simulates the asphalt pavement structure and the cavity disease through a computer, and then obtains a simulated ground penetrating radar image through calculation, and verifies the SSIM-S index through the forward model; the new evaluation index is also inspired by the forward model, so that the SSIM-S index of the ground penetrating radar image quality evaluation method is determined based on the relationship between the forward model and the image quality evaluation index; meanwhile, because the cavity disease characteristics of the experimental group in the forward model are affected by the multi-layer structure of the road, the characteristics are certainly not standard and obvious, the road structure in the control group model only has one layer, which greatly reduces other influencing factors, and the simulated cavity disease characteristics are relatively more standard and obvious; therefore, the image quality of the experimental group model is obtained by calculating the SSIM-S indexes of the two groups of models; Three different forward models of the ground penetrating radar detection of the asphalt pavement cavity disease are constructed; the first layer is a surface layer, and the material is asphalt mixture; the second layer is a base layer, and the material is cement stabilized macadam; the third layer is a roadbed, and the main material is compacted soil and gravel; TA is a ground penetrating radar transmitting antenna, and RA is a radar receiving antenna; when the electromagnetic wave propagates between multiple different media, the distance between the antenna transmitting source and the receiving source will affect the processing result of the final signal, and the burial depth of the target body will affect the detection accuracy, therefore, the influence of the cavity disease with different burial depths and the antenna transmitting-receiving distance on the detection effect of the asphalt pavement cavity disease is analyzed. 2.The method for ground penetrating radar image quality assessment for asphalt pavement void disease identification according to claim 1, characterized in that, The forward model uses a 400MHz Ricker wavelet as an excitation source signal.

3. The ground penetrating radar image quality evaluation method for asphalt pavement cavity disease identification according to claim 2, characterized in that, In the process of acquiring the ground penetrating radar image of the detection road, the antenna transmitting-receiving distance is set to be less than 25cm.

4. The ground penetrating radar image quality evaluation method for asphalt pavement cavity disease identification according to claim 3, characterized in that, In the process of acquiring the ground penetrating radar image of the detection road, the antenna transmitting-receiving distance is set to be [10cm, 20cm].

5. The ground penetrating radar image quality assessment method for asphalt pavement void disease identification according to claim 3 or 4, characterized in that, The antenna transmitting-receiving distance in the forward model is equal to the antenna transmitting-receiving distance set in the process of acquiring the ground penetrating radar image of the detection road. 6.A ground penetrating radar image quality evaluation system for asphalt pavement cavity disease identification, characterized in that, The method comprises the following steps: The ground penetrating radar image acquisition unit: for the detection road, a ground penetrating radar image of the detection road, i.e., a detection B-scan image, is acquired; The control group forward model establishment unit: a corresponding control group forward model is established according to a design file of the detection road, wherein the control group is a single-layer pavement structure obtained by changing a multi-layer asphalt pavement structure in the design file of the detection road; The control group image acquisition unit: based on the established control group forward model, a control group B-scan image is simulated and obtained; The quality evaluation unit: image quality evaluation of the detection B-scan image is performed according to an SSIM-S index; The quality evaluation unit: image quality evaluation of the detection B-scan image is performed according to an SSIM-S index; ； wherein representative image representative image representative image representative image representative image representative image representative image representative image The forward model of the cavity disease of asphalt pavement: The forward model is a simulation method, which firstly simulates the asphalt pavement structure and cavity disease through computer simulation, and then obtains the simulated ground penetrating radar image through calculation, and verifies the SSIM-S index through the forward model; the new evaluation index is also inspired by the forward model, so that the SSIM-S index of the ground penetrating radar image quality evaluation method is determined based on the relationship between the forward model and the image quality evaluation index; at the same time, because the cavity disease characteristics of the experimental group in the forward model are affected by the multi-layer structure of the road, its characteristics are definitely not standard and obvious, and the road structure in the control group model only has one layer structure, which greatly reduces other influencing factors, and the simulated cavity disease characteristics are relatively more standard and obvious; therefore, by calculating the SSIM-S indexes of the two groups of models, the image quality of the experimental group model is obtained; Three different forward models of asphalt pavement cavity disease detection by ground penetrating radar are constructed; the first layer is the surface layer, and the material is asphalt mixture; the second layer is the base layer, and the material is cement stabilized gravel; the third layer is the roadbed, and the main material is compacted soil and gravel; TA is the ground penetrating radar transmitting antenna, and RA is the radar receiving antenna; when the electromagnetic wave propagates between multiple different media, the distance between the antenna transmitting and receiving sources will affect the final signal processing result, and the burial depth of the target body will affect the detection accuracy, therefore, the influence of different buried cavity diseases and different antenna transmitting and receiving distances on the detection effect of asphalt pavement cavity disease is analyzed.

7. The ground penetrating radar image quality evaluation system for asphalt pavement void disease identification of claim 6, wherein, The forward model uses a 400MHz Ricker wavelet as the excitation source signal. 8.The ground penetrating radar image quality evaluation system for asphalt pavement void disease identification of claim 7, wherein, In the process of obtaining the ground penetrating radar image of the detected road, the antenna transmitting and receiving distance is set to be less than 25cm. 9.The ground penetrating radar image quality evaluation system for asphalt pavement void disease identification of claim 8, wherein, In the process of obtaining the ground penetrating radar image of the detected road, the antenna transmitting and receiving distance is set to be [10cm, 20cm].

10. The ground penetrating radar image quality evaluation system for asphalt pavement void disease identification according to claim 8 or 9, characterized in that, The antenna transmitting and receiving distance in the forward model is equal to the antenna transmitting and receiving distance set in the process of obtaining the ground penetrating radar image of the detected road.

Images