CN116266215A - Geometric feature extraction method for pavement recessive disease area - Google Patents
Geometric feature extraction method for pavement recessive disease area Download PDFInfo
- Publication number
- CN116266215A CN116266215A CN202210914878.0A CN202210914878A CN116266215A CN 116266215 A CN116266215 A CN 116266215A CN 202210914878 A CN202210914878 A CN 202210914878A CN 116266215 A CN116266215 A CN 116266215A
- Authority
- CN
- China
- Prior art keywords
- function
- value
- disease
- pavement
- energy
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 201000010099 disease Diseases 0.000 title claims abstract description 69
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 title claims abstract description 69
- 238000000605 extraction Methods 0.000 title claims abstract description 19
- 238000001228 spectrum Methods 0.000 claims abstract description 37
- 239000011800 void material Substances 0.000 claims abstract description 31
- 230000009467 reduction Effects 0.000 claims abstract description 20
- 238000012545 processing Methods 0.000 claims abstract description 11
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 9
- 230000009466 transformation Effects 0.000 claims abstract description 9
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 238000001914 filtration Methods 0.000 claims abstract description 6
- 238000000034 method Methods 0.000 claims description 27
- 238000004364 calculation method Methods 0.000 claims description 9
- 230000002829 reductive effect Effects 0.000 claims description 5
- 238000005070 sampling Methods 0.000 claims description 5
- 238000012790 confirmation Methods 0.000 claims description 4
- 238000012937 correction Methods 0.000 claims description 4
- 230000003068 static effect Effects 0.000 claims description 4
- 230000000670 limiting effect Effects 0.000 claims description 2
- 230000003313 weakening effect Effects 0.000 claims description 2
- 230000007547 defect Effects 0.000 claims 8
- 238000004422 calculation algorithm Methods 0.000 abstract description 13
- 239000004568 cement Substances 0.000 abstract description 8
- 238000012423 maintenance Methods 0.000 abstract description 4
- 238000001514 detection method Methods 0.000 description 7
- 230000000149 penetrating effect Effects 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000003902 lesion Effects 0.000 description 3
- 230000001052 transient effect Effects 0.000 description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 3
- 229910000831 Steel Inorganic materials 0.000 description 2
- 230000001154 acute effect Effects 0.000 description 2
- 238000011088 calibration curve Methods 0.000 description 2
- 230000003247 decreasing effect Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 239000010959 steel Substances 0.000 description 2
- 241000826860 Trapezium Species 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 238000011946 reduction process Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B15/00—Measuring arrangements characterised by the use of electromagnetic waves or particle radiation, e.g. by the use of microwaves, X-rays, gamma rays or electrons
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/885—Radar or analogous systems specially adapted for specific applications for ground probing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Physics & Mathematics (AREA)
- Electromagnetism (AREA)
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- General Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention discloses a cement pavement void geometric feature extraction algorithm based on a CWT energy spectrum. 1. Data acquisition, forward and backward modeling is carried out to obtain original data; 2. preprocessing the original data; 3. performing continuous wavelet transformation processing on the preprocessed signals to reconstruct an energy spectrum; 4. performing dimension reduction on the energy spectrum, performing discrete wavelet decomposition on the dimension reduction data, and removing interference; 5. solving an energy function probability density function of a disease area, and solving an expected value of the function; 6. setting a threshold value by using an expected value, filtering out extreme points generated by interference, and solving two nearest minimum value points on the left and right sides of the maximum value point to determine the boundary width of the disease; 7. and calculating an improved generalized S transformation spectrum of the disease area, and solving a maximum value with minimum index in a dimension reduction function after dimension reduction, so as to calculate the depth. The problem that the existing radar map identification cannot accurately position a disease area and cannot determine the geometric characteristics of a void disease area is solved, and a basis is provided for pavement maintenance.
Description
Technical Field
The invention belongs to the field of pavement maintenance, and particularly relates to a method for reconstructing and extracting geometric features of a disease area by utilizing a ground penetrating radar signal continuous wavelet transformation energy spectrum.
Background
Cement pavement plays an important role in the traffic network in China, has high bearing capacity, good stability and low maintenance cost, and is still a road surface structure of a trunk road (province road and county road) which is widely applied. Local gaps are generated between the panels and the base layer due to uneven construction, uneven settlement of the roadbed, influence of vehicle load and temperature stress, and the like. After caulking materials between the plates fall off, rainwater easily enters the void area, the range of the water-containing void area can be further enlarged under the action of alternating load, the development of diseases in the void area is aggravated, the influence on the bearing capacity of a road is large, the problems of settlement and staggering of cement plates are caused, and finally, broken plates are generated, so that traffic is seriously influenced. The method for determining the geometric dimension of the damaged area is urgently needed to be established, so that the bearing capacity of the pavement slab is evaluated, the pavement is maintained before the slab is broken, the service life of the pavement is prolonged, and a scientific basis is provided for accurate pavement maintenance.
The method for detecting the void diseases mainly comprises a pulse response method, a sound vibration method, a vibration sensing method, a Ground Penetrating Radar (GPR) and the like, wherein the ground penetrating radar is the most effective nondestructive detection technology in the pavement void detection, and is widely applied to pavement diseases and thickness of structural layers. The A-Scan signal of GPR is the minimum information unit of GPR, and also is the basic element formed by B-Scan map, and contains information such as depth, medium characteristics, thickness dimension and the like, so that if disease identification can be performed by using the GPR signal, the disease boundary position can be accurately positioned. The existing GPR pavement detection method has the problems that the working efficiency is low and the subjectivity of the identification result is strong because a large amount of data generated during the existing GPR pavement detection is dependent on whether diseases and positions of the diseases exist or not through data processing by experienced professionals.
Echo signals of Ground Penetrating Radar (GPR) are transient non-stationary signals, and the distribution of different frequency components in the signals on a time axis cannot be revealed by the traditional Fourier transform due to the characteristics of a basis function. The wavelet transform (CWT) is a time-frequency localization analysis method with a fixed window size and a changeable shape, and is used for capturing local and whole characteristics of signals through different scales, and is often used for time-frequency analysis of signals or accurate transient positioning of signals, especially for signals with abrupt transient frequencies. Therefore, the continuous wavelet transformation provides a new thought and method for identifying the ground penetrating radar map, positioning the disease area and determining the geometric characteristics of the disease area.
The patent No. CN109444176A provides a method for detecting the concrete void depth under a steel shell, which considers the influence of the water content of the concrete on the void depth by establishing a void depth calibration curve, adopts different calibration curves aiming at different steel plate thicknesses and concrete water content, but does not extract geometric parameters of the void width. The patent No. CN110487910A provides a panel dam panel void and positioning detection method based on vibration sensing technology, which is used for positioning the void position by detecting the amplitude and acceleration variation of each sensor on the panel. The above patent is not directed to GPR signals, but cannot solve the problem of difficult GPR signal interpretation.
At present, geometric feature parameter extraction of a disease area is lacking, and the applicant starts with continuous wavelet transformation, and proposes a cement pavement void geometric feature extraction method based on reconstructed CWT energy spectrum, so that transverse dimension extraction of the void area is realized, and geometric parameters of the disease area are provided for pavement structure safety evaluation.
Disclosure of Invention
In order to solve the problem that the existing radar spectrum identification method cannot accurately position a disease region and cannot determine geometric feature parameters of a void disease region, the method for extracting the geometric feature of the disease region based on reconstructing a CWT energy spectrum of a GPR original signal is provided.
The invention is realized by the following steps.
And 2, preprocessing the original GPR data (static correction, gain, background removal and F-K offset) to obtain a preprocessed signal.
And 3, performing Continuous Wavelet Transform (CWT) processing on the preprocessed signals, and then constructing a reconstructed energy spectrum. Compared with the original B-Scan, the reconstructed energy spectrum can more effectively reveal the transverse and longitudinal distribution of different disease areas, and can judge the medium difference of different areas.
And 4, performing dimension reduction on the energy spectrum to obtain a dimension reduction curve, performing discrete wavelet decomposition on the dimension reduction curve by adopting a sym4 wavelet basis function, removing background noise, and improving the smoothness of the curve.
And 5, obtaining an energy function probability density function of the dimension reduction curve, and using an expected value of the function as a threshold setting condition for obtaining an extreme point and a confirmation width.
Where μ, σ are the mean and variance, respectively, of the reduced dimension data matrix.
And 6, extracting the void width, setting a threshold value by using the expected values in the steps 4 and 5, filtering out extreme points generated by interference, solving two nearest minimum value points on the left and right of the maximum value point to determine the boundary width of the disease, and storing the widths of all disease areas in a vector L.
And 7, obtaining depth according to the principle of improving the maximum value of the generalized S transformation spectrum. Calculating the improved S-transform spectrum Gs of the ith channel A-Scan data in the disease area (the area corresponding to the non-zero element) by the non-zero element value in the width vector i (f, t) for which spectrum Gs i Dimension reduction is performed to obtain (time-energy)) Function g i (t), function g i (t) is a reduced-dimension energy function that improves the S-transform; solving for a function g i And (3) the extreme point m with the smallest index in the largest n values in the (t) is the depth information.
g i (t n )＝max(Gs i (f，t n ))，n＝1，2，…，N (3)
g i (t)＝[g i (t 1 )，…，g i (t N )] (4)
Where N is the total sampling point number of the ith track A-Scan data, g i (t n ) Is Gs i At t n Maximum value of time.
Preferably, when discrete wavelet decomposition is performed in the step 4, the principle of decomposing the layer number is that when the number of extreme points of which the function is lower than the average value after dimension reduction is less than half of the total number of extreme points, the layer number is 3, and conversely, the dimension reduction layer number is 4.
Preferably, when two minimum value points closest to the maximum value point are obtained in the step 6 to determine the boundary width of the disease, since the radar waves similar to the disease have overlapping phenomenon, the hyperbolic characteristic of the disease is intersected, and further, part of the minimum value points cannot reflect the actual width of the disease, so that a relative height factor alpha, namely, the height difference of the value of the minimum value point relative to the minimum value in the two maximum values of the left and right of the minimum value point is introduced. When the relative height exceeds a set threshold, it is necessary to filter out interference from multiple local minima points.
α＝min(M 1 ，M 2 )-y min (5)
Wherein y is min Is minimum value, M 1 ，M 2 Is two maxima around the minimum.
Preferably, in step 6, the number of tracks is used as a constraint for the case where there is no minimum point at the maximum point boundary.
Preferably, in step 7, the maximum energy is not the original Ricker sub-peak due to superposition of the sub-waves, and therefore, the extreme point with the smallest index out of the n maximum values in the function is obtained as the depth calculation.
PreferablyIn the step 7, the disease depth calculation formula is
Where t=m·dt, ε is the dielectric constant of the concrete.
The invention has the beneficial effects that
(1) The method is based on the GPR minimum signal unit A-scan to obtain the width and depth dimensions of the void region, and compared with the existing method which utilizes a GPR B-scan map (a map formed by overlapping A-scan along a measuring line) to obtain the dimension of the disease region, the method has higher recognition precision and can effectively determine the boundary of the disease region.
(2) The CWT technology of signal processing is used for GPR signal processing, the characteristic that the disease area has energy concentration or strong reflection is utilized, the CWT energy spectrum of the reconstructed GPR signal is provided, and the disease width and depth size are calculated by combining the dimension reduction idea. The method for extracting the geometric characteristic parameters of the diseases does not need a large number of samples for modeling, can conveniently combine the size information of the diseases with a finite element analysis tool, such as ANSYS and ABQUAS software, and provides technical parameters for evaluating the bearing capacity of the pavement slab.
(3) The geometric feature extraction method for the cement pavement void disease area can reduce the participation of manpower in void identification and reduce the interference of the artificial experience on the identification result.
Drawings
FIG. 1 is a general flow chart of the present invention;
FIG. 2 is a forward modeling model A;
FIG. 3 is a forward modeling model B;
FIG. 4 is a CWT result of a normal lane;
fig. 5 is a CWT result of the de-lanes;
FIG. 6 is a schematic diagram of a process of reconstructing an energy spectrum;
FIG. 7 is a reconstructed energy spectrum of model A;
FIG. 8 is a reconstructed energy spectrum of model B;
FIG. 9 is a dimension reduction signal DWT decomposition of model B;
FIG. 10 is a flowchart of a disease point extraction algorithm;
FIG. 11 is a schematic diagram of model void width determination;
FIG. 12 is a maximum value schematic diagram with minimum index;
FIG. 13 is a flow chart for determining depth of disease;
FIG. 14 is a depth and width extraction result of model A;
FIG. 15 is a depth and width extraction result for model B;
FIG. 16 is a histogram of width extraction errors for model A;
FIG. 17 is a histogram of width extraction errors for model B;
FIG. 18 is an inverted indoor model;
FIG. 19 is an energy spectrum of an indoor model reconstruction;
FIG. 20 is a graph showing the width and depth extraction results of an indoor model;
fig. 21 is a histogram of width extraction errors of an indoor model.
The specific embodiment is as follows:
And 2, preprocessing the original GPR data, wherein the preprocessing mainly comprises extracting an average channel, static correction, gain, background removal, band-pass filtering, and then weakening the hyperbolic characteristic of the isolated body through F-K offset processing to obtain a GPR preprocessing signal.
And 3, reconstructing an energy spectrum, wherein the specific operation flow is shown in fig. 6. A-scan signal x for a pre-processed GPR signal i (t) performing Continuous Wavelet Transform (CWT) processing, the processed normal and de-lane results are shown in fig. 4,5, and the de-lane CWT results are shown as energy concentration areas. Its wavelet coefficient spectrum is G i (f, t) determining the maximum energy of CWT
The corresponding frequency is used as the main frequency of the signal>
Then find +.>
Corresponding time-energy signal y i And (t) traversing all the A-scan data of the pre-processed signal to obtain a reconstructed energy spectrum B (x, t) (shown in FIG. 7 and FIG. 8).
And 4, in order to obtain the disease size parameter, the dimension of the two-dimensional matrix B (x, t) is required to be reduced, so that an energy function y (x) is obtained, at the moment, the local convex peak of the disease can become a local maximum point of the curve y, and the information of the maximum point can be further used for determining the transverse position of the disease, wherein the dimension reduction process is shown in a formula (6).
y(x i )＝B max (x i ，t)i＝1，2，3…N (6)
Wherein y (x) i ) At x for energy spectrum B i Maximum value at, N is the maximum sampling channel number, y (x i ) Is the i-th component constituting the vector y (x). The method comprises the following steps:
y(x)＝[y(x 1 )，y(x 2 )，…，y(x N )] (7)
due to the introduction of background noise caused by the dimension reduction, the smoothness of the interference curve, the signal y (x) is subjected to 4-layer discrete wavelet decomposition by adopting sym4 wavelet basis functions subjected to periodic boundary processing (fig. 9).
Wherein cD i For high frequency wavelet coefficients, cA n Is a low frequency wavelet coefficient. n is the number of decomposition layers, the smaller n is, the lower the frequency band resolution is, and on the contrary, the higher the frequency band resolution is. The decomposed approximation gamma (x) signal does not affect the main maximum point position of the curve y (x), and each sample in the function gamma (x) is valued gamma (x) i ) The energy range of the sample is divided, and the sample energy distribution is approximately subjected to lognormal distribution through actual statistical analysis and combination of the positive and negative properties of the sample skewness. Therefore, the concentration of the sample energy distribution is described based on the expected value E of the sample energy distribution, and the obtained concentration can be used as a threshold condition for obtaining the extreme point and the confirmation width.
And 5, extracting the void width.
(1) Knowing the B (x, T) time domain energy profile T (T), reconstructing the energy to obtain a reconstructed energy spectrum.
(2) Estimating each sample gamma (x) in an energy function gamma (x) of the reconstructed energy spectrum i ) The expected value is noted as K.
(3) Solving all maximum points of the function gamma (x) and the values thereof to form a vector A 1×N (N is the number of tracks of the A-scan signal in the GPR data), if A 1×N Element a of (2) i ＜A th Then a (i) =0, threshold a th The value range of K1 is set to be 80% -90% of the calculated result K.
(4) Considering the interference between disease areas, the method is required to further reject the product of A 1×N Local minimum points of the function phi (x) reconstructed by the medium maximum points are recorded as H by the average value of the relative heights of all the minimum points of phi (x) mean If phi (x) i )＞H mean Let ψ (i) =0. The vector psi (m) stores the extreme point index in A, and m is the number of all the extreme points.
(5) Let vector L 1×N =0, and the boundary width is determined by calculating two nearest minimum points on the left and right of the maximum point in the function y (x) from the index of the extremum point. And for the situation that the boundary has no minimum value point, the number of tracks is required to be used as a limiting condition, and a disease point identification algorithm flow chart is shown in fig. 10.
The input parameter disease point is the result of the algorithm calculation of FIG. 10, where the function P min ，R min The left and right minimum value points (boundary points) on both sides of the maximum value point (disease point) are obtained. The return value is an index of a disease boundary point or False, K is an energy drop threshold, and the width is obtained so as to prevent overlapping of left and right extreme points of adjacent disease points. f (f) 1 ，f 2 The calculated left and right boundaries of the lesion area, the vector L stores the widths of all the lesion areas, and the lesion width extraction algorithm is shown in fig. 11.
And 6, extracting the emptying depth.
Knowing the calculated width vector L 1×N And the preprocessed data B (x, t),according to the non-zero element value in the width vector, calculating an improved S transformation spectrum Gs of the ith channel A-Scan data in the disease area (the area corresponding to the non-zero element) i (f, t) for which spectrum Gs i Dimension reduction is carried out to obtain a (time-energy) function g i (t), function g i (t) is a reduced-dimension energy function that improves the S-transform; solving for a function g i And (3) the extreme point p with the smallest index in the largest n values in the (t) is the depth information. Because the superposition of the sub-waves results in that the maximum energy is not the original Ricker sub-peak, the extreme point with the minimum index in the extreme points is obtained to be used as depth calculation, for example, the extreme point with the minimum index in three values with the maximum extreme point number 2 is recorded as the depth of the channel A-Scan, the principle of the maximum value with the minimum index is shown in fig. 12, and the emptying depth extraction algorithm flow chart is shown in fig. 13.
g i (t n )＝max(Gs i (f，t n ))，n＝1，2，…，N (9)
g i (t)＝[g i (t 1 )，…，g i (t N )] (10)
Where N is the total sampling point number of the ith track A-Scan data, g i (t n ) Is Gs i At t n Maximum value of time.
The forward modeling models a, B were processed according to the algorithm described above, fig. 14, fig. 15 are depth and width extraction results of models a and B, fig. 16, and fig. 17 are graphs of the models a and B and error bars. It can be seen that the information of the basic disease points is effectively identified, but misjudgment exists. The misjudgment at the position 2 in the model A is caused by noise interference caused by pretreatment and F-K treatment; and the model B has no misjudgment, and all disease points are effectively identified, so that the algorithm can be effectively used for identifying the void disease. In the width error bar graph, detection is the calculated value of the proposed algorithm. Fig. 16 shows the width errors of different shapes, and it is obvious that the targets (3-5) with triangles have larger width errors and the round targets (1, 2) have smaller errors. The reason is that the edge acute angle characteristics of the right triangle, the isosceles triangle and the isosceles trapezoid are that the longitudinal resolution at the acute angle is poor, so that the reflected echo energy is small, and the effective width is reduced; the middle area of the triangle is larger in longitudinal dimension, so that the reflected energy is stronger and the error is smaller, but the calculation error for the upper edge width of the isosceles trapezium (6, 7) is smaller. Compared with a triangle, under the same width, the effective width of the circular middle area is larger due to the special structure of the round hole, the boundary of the round hole is gentle, and therefore, the error is smaller than that of a triangle. Fig. 17 shows a rectangular disease area width error of about 5% and a circular error of about 10%, with the overall smaller than circular and the disease error deeper. And the shape of the actual void area is many rectangular and circular. Thus, the algorithm presented herein can effectively perform better width calculations for regions containing rectangular and circular features.
And 7, in order to further verify the correctness and the accuracy of the method provided by the invention, constructing indoor emptying models with the same shape, different sizes and different depths. As shown in fig. 18, the model length x width x thickness is 2.07 x 0.4 x 0.5m, 10 void round holes with 15cm intervals are designed, the diameters are sequentially 100, 90-30, 25 and 16mm, the positions of all circle centers are positioned on a central line of 20cm, the model material is cement concrete, the test is performed by adopting an US radar 1G antenna, the sampling frequency is 16GHz, the time window is 11.675ns, the total channel number is 207, and the original GPR data are obtained.
FIG. 19 is a three-dimensional visualization of the reconstructed energy spectrum of an indoor build model. From the figure, it can be seen that 1-9 are voids, the energy of which increases progressively in sequence, and the energy of the deeper voids is weaker. Fig. 20 shows the calculation result of the width and depth processing algorithm, except that the energy of the cavity 10 is too small to be resolved from the background threshold, the rest is correctly identified, the energy is consistent with the trend of fig. 14, the depth position is distributed at the lower edge of the black dot, and the actual depth is based on the center of the black dot, so that the proposed algorithm can effectively estimate the depth position of each cavity. FIG. 21 is a bar graph of the width error of the void disease, the Detection is a calculated value of the proposed algorithm, the hole error is larger with deeper depth, but the error basically shows a decreasing trend with decreasing depth. Therefore, the method can effectively solve the problem of calculation of the width and estimation of the depth of the void area and judge the safety of the pavement structure.
The above is only for illustrating the technical idea of the present invention, and the protection scope of the present invention is not limited by this, and any modification made on the basis of the technical scheme according to the technical idea of the present invention falls within the protection scope of the claims of the present invention.
Claims (9)
1. The geometrical feature extraction method of the pavement recessive disease area is characterized by comprising the following steps of:
step one, data acquisition, simulating a model by using gprMax or performing actual data acquisition by using GPR to obtain original GPR data;
preprocessing the original GPR data (static correction, gain, background removal and F-K offset) to obtain a preprocessed signal;
and thirdly, performing Continuous Wavelet Transform (CWT) processing on the preprocessed signals, and then constructing a reconstructed energy spectrum. Compared with the original B-Scan, the reconstructed energy spectrum can more effectively reveal the transverse and longitudinal distribution of different disease areas, and can judge the medium difference of different areas;
step four, reducing the dimension of the energy spectrum to obtain a dimension-reduced curve, performing discrete wavelet decomposition on the dimension-reduced curve by adopting a sym4 wavelet basis function, removing background noise and improving the smoothness of the curve;
step five, solving an energy function probability density function of the dimension reduction curve, and utilizing an expected value of the function as a threshold setting condition for solving an extreme point and a confirmation width;
step six, extracting the void width, setting a threshold value by using the expected value in the step four and the step five, filtering out extreme points generated by interference, solving two nearest minimum value points on the left and right of the maximum value point to determine the boundary width of the disease, and storing the widths of all disease areas in a vector L;
seventh, according to the maximum principle of the improved generalized S transformation spectrum, the depth and the non-zero element value in the width vector are obtained, and the improved S transformation spectrum Gs of the ith channel A-Scan data in the disease area (the area corresponding to the non-zero element) is calculated i (f, t) for which spectrum Gs i Dimension reduction is carried out to obtain a (time-energy) function g i (t), function g i (t) is a reduced-dimension energy function that improves the S-transform; solving for a function g i The extreme point m with the smallest index in the largest n values in the (t) is the depth information;
g i (t n )＝max(Gs i (f，t n ))，n＝1，2，…，N
g i (t)＝[g i (t 1 )，…，g i (t N )]
where N is the total sampling point number of the ith track A-Scan data, g i (t n ) Is Gs i At t n Maximum value of time.
2. The method of claim 1, wherein the preprocessing of the raw GPR data in the second step mainly comprises extracting an average channel, static correction, gain, background removal, bandpass filtering, and then weakening the isolated hyperbolic characteristic by F-K offset processing.
3. The method for extracting geometric features of a pavement recessive disease area according to claim 1, wherein the method for dimension reduction of the energy spectrum in the fourth step is to project the energy spectrum onto a train Number-Magnitude plane, so that a peak value in the energy spectrum is converted into an extreme point on a plane view.
4. The method for extracting geometric features of a pavement recessive disease area according to claim 1, wherein the sym wavelet basis function adopted in the fourth step is subjected to 4 times of discrete wavelet decomposition to remove background noise, wherein the principle of the number of decomposition layers is that when the number of extreme points of the function lower than the average value after dimension reduction is less than half of the number of total extreme points, the number of layers is 3, and conversely, the number of dimension reduction layers is 4.
5. The method for extracting geometric features of a pavement recessive disease area according to claim 1, wherein the expected value E of a probability density function of an energy function of the disease area is obtained in the fifth step as a condition for filtering interference generated by two adjacent disease signals.
6. The method for extracting geometric features of a pavement hidden trouble area according to claim 1, wherein the threshold value K of the confirmation width in the fifth step is 80% -90% of the expected value E.
7. The geometric feature extraction method of a pavement recessive defect region according to claim 1, wherein in the step six, when the defect region width is obtained, the defect point is first identified, then the defect width is determined by obtaining minimum value points on both sides of the maximum value point, and the number of tracks is required as a limiting condition for the case that the boundary has no minimum value point.
8. The method for extracting geometric features of a pavement hidden defect area according to claim 1, wherein when two minimum value points closest to a maximum value point are obtained in the sixth step to determine the boundary width of the defect, a relative height factor α is required to be introduced, that is, a height difference between the value of the minimum value point and the minimum value of the two maximum values, and when the relative height exceeds a set threshold value, interference of a plurality of local minimum value points is required to be filtered.
9. The method for extracting geometric features of a pavement recessive defect region according to claim 1, wherein in the step seven, when the defect depth is obtained, the maximum energy is not the original Ricker sub-peak due to superposition of sub-waves, so that the extreme point with the minimum index in the maximum n values in the function is obtained as the depth calculation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210914878.0A CN116266215B (en) | 2022-08-01 | 2022-08-01 | Geometric feature extraction method for pavement recessive disease area |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210914878.0A CN116266215B (en) | 2022-08-01 | 2022-08-01 | Geometric feature extraction method for pavement recessive disease area |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116266215A true CN116266215A (en) | 2023-06-20 |
| CN116266215B CN116266215B (en) | 2024-01-26 |
Family
ID=86744116
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210914878.0A Active CN116266215B (en) | 2022-08-01 | 2022-08-01 | Geometric feature extraction method for pavement recessive disease area |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116266215B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116973914A (en) * | 2023-09-06 | 2023-10-31 | 哈尔滨工业大学 | Road hidden disease three-dimensional reconstruction method based on three-dimensional ground penetrating radar |
| CN117077450A (en) * | 2023-10-17 | 2023-11-17 | 深圳市城市交通规划设计研究中心股份有限公司 | Road void area volume evolution prediction method, electronic equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101303802A (en) * | 2008-06-26 | 2008-11-12 | 吉林大学 | Method and system for on-line automatically beforehand judgment of overloading wagon |
| CN101751782A (en) * | 2009-12-30 | 2010-06-23 | 北京大学深圳研究生院 | Crossroad traffic event automatic detection system based on multi-source information fusion |
| CN102509291A (en) * | 2011-10-31 | 2012-06-20 | 东南大学 | Pavement disease detecting and recognizing method based on wireless online video sensor |
| US20190247662A1 (en) * | 2017-12-04 | 2019-08-15 | Neuroenhancement Lab, LLC | Method and apparatus for neuroenhancement to facilitate learning and performance |
| CN111749092A (en) * | 2020-06-02 | 2020-10-09 | 交通运输部公路科学研究所 | Porous asphalt pavement flying disease detection method based on noise signal identification |
-
2022
- 2022-08-01 CN CN202210914878.0A patent/CN116266215B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101303802A (en) * | 2008-06-26 | 2008-11-12 | 吉林大学 | Method and system for on-line automatically beforehand judgment of overloading wagon |
| CN101751782A (en) * | 2009-12-30 | 2010-06-23 | 北京大学深圳研究生院 | Crossroad traffic event automatic detection system based on multi-source information fusion |
| CN102509291A (en) * | 2011-10-31 | 2012-06-20 | 东南大学 | Pavement disease detecting and recognizing method based on wireless online video sensor |
| US20190247662A1 (en) * | 2017-12-04 | 2019-08-15 | Neuroenhancement Lab, LLC | Method and apparatus for neuroenhancement to facilitate learning and performance |
| CN111749092A (en) * | 2020-06-02 | 2020-10-09 | 交通运输部公路科学研究所 | Porous asphalt pavement flying disease detection method based on noise signal identification |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116973914A (en) * | 2023-09-06 | 2023-10-31 | 哈尔滨工业大学 | Road hidden disease three-dimensional reconstruction method based on three-dimensional ground penetrating radar |
| CN116973914B (en) * | 2023-09-06 | 2024-06-11 | 哈尔滨工业大学 | Road hidden disease three-dimensional reconstruction method based on three-dimensional ground penetrating radar |
| CN117077450A (en) * | 2023-10-17 | 2023-11-17 | 深圳市城市交通规划设计研究中心股份有限公司 | Road void area volume evolution prediction method, electronic equipment and storage medium |
| CN117077450B (en) * | 2023-10-17 | 2024-03-26 | 深圳市城市交通规划设计研究中心股份有限公司 | Road void area volume evolution prediction method, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116266215B (en) | 2024-01-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112462346B (en) | Ground penetrating radar subgrade disease target detection method based on convolutional neural network | |
| CN116266215B (en) | Geometric feature extraction method for pavement recessive disease area | |
| CN111476088B (en) | Asphalt pavement water damage identification model construction method, identification method and system | |
| CN109782274B (en) | Water damage identification method based on time-frequency statistical characteristics of ground penetrating radar signals | |
| Bai et al. | Comparison of alternatives to amplitude thresholding for onset detection of acoustic emission signals | |
| CN100535955C (en) | Method for recognizing outlier traffic data | |
| CN110208806B (en) | Marine radar image rainfall identification method | |
| CN112666554A (en) | Method for identifying radar amplitude characteristic crack width of asphalt pavement | |
| CN110147716B (en) | SAR image internal wave region detection method based on combination of frequency domain and airspace | |
| CN102621531B (en) | Rainfall interference suppression method based on X-band radar images | |
| CN112818924B (en) | Road deep disease early warning method based on association rule analysis | |
| CN114675261B (en) | Urban underground disease body full life cycle management method based on ground penetrating radar | |
| CN114775382B (en) | Ultrasonic-based road surface quality detection method for highway traffic engineering | |
| CN106338729A (en) | Method for inversing railway base track bed contamination rate by using ground penetrating radar | |
| CN113009447A (en) | Road underground cavity detection early warning method based on deep learning and ground penetrating radar | |
| CN115390033A (en) | Method, system and device for detecting cavity and water loss under road surface based on ground penetrating radar | |
| CN116973914B (en) | Road hidden disease three-dimensional reconstruction method based on three-dimensional ground penetrating radar | |
| CN109001728B (en) | Method and device for detecting internal diseases of reinforced concrete based on ground penetrating radar | |
| CN1865983A (en) | Fractal elimination method for mass data in ultrasonic intelligent detection | |
| Guo et al. | Research on tunnel lining image target recognition method based on YOLOv3 | |
| CN116626685B (en) | River sediment real-time monitoring method and system based on machine learning | |
| CN116203557B (en) | Traditional stone wall internal damage and anomaly nondestructive identification method based on ground penetrating radar | |
| CN117077448B (en) | Road void area evolution prediction method, electronic equipment and storage medium | |
| CN113837163B (en) | Tunnel monitoring method and system based on three-dimensional ground penetrating radar and storage medium | |
| CN117518281B (en) | Method for quantitatively evaluating pavement hidden diseases based on three-dimensional ground penetrating radar |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |