Forward analysis method, system and terminal for dielectric characteristics of mixed medium ground penetrating radar
Through the random shuffling algorithm combined with the time domain finite difference method, a fine grid model of mixed medium was constructed, which solved the electrical parameter incompatibility and electromagnetic wave diffraction problems of mixed medium ground-penetrating radar simulation in the existing technology, and realized the GPR simulation simulation Stability and effectiveness provide accurate analysis of the dielectric constant distribution.
Patent Information
- Application Number
- CN202111595290.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-12-24
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2041-12-24
AI Technical Summary
The existing ground-penetrating radar simulation method has incompatibility of electrical parameters and electromagnetic wave diffraction when processing mixed media, resulting in GPR images that do not match the actual situation, and the calculation speed is slow and the memory requirements are large, making it difficult to build a hybrid medium with unknown electrical parameters. Model.
Using the random shuffling algorithm combined with the time domain finite difference method (FRS-FDTD), the coarse mesh model of the mixed medium is divided into fine mesh models, and the corresponding dielectric parameters are assigned to each fine mesh model, through the MATLAB platform Run the FDTD algorithm to generate GPR images, and analyze the media signals using image entropy theory and wavelet transform mode maximum method.
The effectiveness and stability of GPR simulation of mixed media are realized, electromagnetic attenuation and energy dissipation are suppressed, dielectric constant distribution is mastered, and theoretical basis for GPR detection parameter correction and image interpretation is provided.
Smart Images

Figure CN114880894B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of ground penetrating radar, and more specifically, to a forward modeling analysis method, system and terminal for dielectric characteristics of a mixed medium ground penetrating radar. Background Art
[0002] Ground Penetrating Radar (GPR) is an advanced non-destructive testing technology that has been widely used in engineering testing fields such as roads, bridges, and underground tunnels. The detection results of GPR are closely related to the dielectric properties of the target medium, such as conductivity, magnetic permeability, and dielectric constant. The uncertainty of electrical parameters after multi-media mixing will directly affect the GPR imaging effect. Therefore, using GPR simulation technology to pre-master the electromagnetic response characteristics and dielectric properties of the mixed medium is of great significance to the imaging, processing, and interpretation of the actual GPR acquisition signal.
[0003] Commonly used GPR simulation modeling methods include the time-domain finite element method and the time-domain finite difference method. The time-domain finite element method FETD uses free triangular mesh partitioning, which has the characteristics of flexible, natural and effective construction of mesh models. Its disadvantages are slow calculation speed, large memory requirements, and severe electromagnetic scattering. It is not suitable for building mixed medium models with unknown electrical parameters. The time-domain finite difference method FDTD has the advantages of direct time-domain calculation, simple programming, and saving storage space. By assigning electrical parameters to Yee cells in the grid space, it can handle problems such as electromagnetic scattering of complex geometric shapes and inhomogeneous media. However, the limitation of the FDTD algorithm is that corresponding mesh models need to be constructed for different field source environments. The main reason is that mesh models with different electrical properties are easily affected by electrical incompatibility and electromagnetic wave diffraction, resulting in the GPR images simulated by FDTD being inconsistent with the actual situation.
[0004] Therefore, how to study and design a forward analysis method, system and terminal for the dielectric properties of mixed medium ground penetrating radar that can overcome the above-mentioned defects is a problem that we urgently need to solve. Summary of the invention
[0005] In order to solve the deficiencies in the prior art, the purpose of the present invention is to provide a forward analysis method, system and terminal for the dielectric properties of mixed medium ground penetrating radar, and to construct a mixed medium fine grid model by combining a random shuffle algorithm with a finite difference time domain method (FRS-FDTD) to carry out GPR forward simulation, which is effective and stable and realizes the GPR simulation of mixed media.
[0006] The above technical objectives of the present invention are achieved through the following technical solutions:
[0007] In the first aspect, a forward analysis method for dielectric properties of a mixed medium ground penetrating radar is provided, comprising the following steps:
[0008] The coarse grid model of mixed medium established based on FDTD is divided into several fine grid models that meet the boundary conditions, and the corresponding dielectric parameters are assigned to the Yee cells in each fine grid model.
[0009] The random shuffle algorithm is used to randomly shuffle all Yee cells with equal probability, and the shuffled Yee cell array is reassembled into a coarse grid model in the form of a differential grid;
[0010] Program the FDTD algorithm on the MATLAB language platform to generate a geoelectric model of the mixed medium and obtain the GPR image under the pulse electromagnetic wave excitation state;
[0011] Image entropy theory is used to quantitatively describe the entropy value of GPR images in mixed media, and waveform analysis method and wavelet transform modulus maximum method are used to analyze the GPR single-channel signal of mixed media.
[0012] Furthermore, the Yee cell assignment process in the fine grid model is specifically as follows:
[0013] Determine the number of mixed media and dielectric parameters in the fine grid model;
[0014] Assign corresponding dielectric parameters to each mixed medium.
[0015] Furthermore, the dielectric parameters include electrical conductivity, magnetic permeability and dielectric constant.
[0016] Furthermore, the subdivision process of the fine grid model is specifically as follows:
[0017] Under the condition that the solution of the finite difference time domain equation system has stable convergence, the restrictive conditions of the space step and time step of the grid model are defined;
[0018] The shape, size and dielectric properties of the mixed medium are determined according to restrictive conditions to achieve the minimum division of the fine grid model.
[0019] Furthermore, the calculation formula of the restrictive condition is specifically:
[0020] The necessary condition for the spatial step size is:
[0021]
[0022] The stability condition for the time step is:
[0023]
[0024] Among them, Δx and Δy are the spatial steps in the x and y directions respectively; Δt is the time step; c is the propagation speed of electromagnetic waves in the air, 0.3m / ns; λ represents the transmission wavelength of the electromagnetic wave.
[0025] Furthermore, the process of randomly scrambling the Yee cells with equal probability is specifically as follows:
[0026] The specific shapes of various mixed media are placed into a rectangular box with a perfect matching layer as the boundary absorption condition;
[0027] The random shuffling algorithm is used to randomly sort the shape model to obtain the scrambled matrix coordinates of the scrambled shape model.
[0028] Furthermore, the acquisition process of the GPR image is specifically as follows:
[0029] The matrix position information generated by rearranging all shapes is used to generate the plaintext image sequence A row by row, and the scrambled matrix R is obtained;
[0030] The FDTD method is combined to carry out forward simulation imaging of ground penetrating radar on the mixed medium to obtain the GPR image.
[0031] Secondly, a forward modeling analysis system for dielectric properties of mixed medium ground penetrating radar is provided, including:
[0032] The model partitioning module is used to partition the FDTD coarse grid model into several fine grid models that meet the boundary conditions, and assign corresponding dielectric parameters to the Yee cells in each fine grid model;
[0033] The random scrambling module is used to use the random shuffling algorithm to randomly scramble all Yee cells with equal probability, and reassemble the scrambled Yee cell array into a coarse grid model in the form of a differential grid;
[0034] The image acquisition module is used to program and run the FDTD algorithm program on the MATLAB language platform to generate a geoelectric model of the mixed medium and obtain the GPR image under the pulse electromagnetic wave excitation state;
[0035] The characteristic analysis module is used to quantitatively describe the entropy value of the mixed medium GPR image using the image entropy theory, and to analyze the GPR single-channel signal of the mixed medium using the waveform analysis method and the wavelet transform modulus maximum method.
[0036] In a third aspect, a computer terminal is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the forward analysis method for dielectric properties of a mixed medium ground penetrating radar as described in any one of the first aspects is implemented.
[0037] In a fourth aspect, a computer-readable medium is provided, on which a computer program is stored, and the computer program is executed by a processor to implement the forward analysis method of dielectric properties of a mixed medium ground penetrating radar as described in any one of the first aspects.
[0038] Compared with the prior art, the present invention has the following beneficial effects:
[0039] The forward analysis method of dielectric properties of mixed medium ground penetrating radar proposed in the present invention uses a random shuffling algorithm combined with a finite-difference time-domain method (FRS-FDTD) to construct a mixed medium fine grid model to carry out GPR forward simulation. It is effective and stable, realizes the GPR simulation of mixed media, effectively suppresses the electromagnetic attenuation and energy dissipation in the mixed medium fine grid model, and uses the wavelet transform modulus maximum method to grasp the distribution of the dielectric constant of the mixed medium. It can serve as a theoretical basis for radar parameter correction and image interpretation when GPR detects mixed media. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of this application, and do not constitute a limitation of the embodiments of the present invention. In the drawings:
[0041] Figure 1 Schematic diagram of the coarse grid and fine grid field distribution based on the FDTD method in an embodiment of the present invention;
[0042] Figure 2 Schematic diagram of a mixed medium fine grid model based on a frequency domain finite difference method in an embodiment of the present invention, (a) is a schematic diagram of field distribution of the mixed medium fine grid model, and (b) is a two-dimensional fine grid distribution of a three-dimensional model;
[0043] Figure 3 is a diagram of the imaging principle of the GPR in an embodiment of the present invention;
[0044] Figure 4 is the geoelectric model of the geological radar in the embodiment of the present invention, (a) is the geoelectric model of the single medium geoelectric model, (b) is the mixed medium geoelectric model;
[0045] Figure 5 1 are GPR forward modeling images of different media types in the embodiment of the present invention, (a) to (d) GPR forward modeling images represent dry sand, dry loam, limestone and mixed media respectively, (e) to (h) are GPR images after signal processing of dry sand, dry loam, limestone and mixed media;
[0046] Figure 6 is a GPR image of a filling medium in an embodiment of the present invention;
[0047] Figure 7 is the image entropy of different filling media in the embodiment of the present invention;
[0048] Figure 8 1 is a waveform diagram of a single-channel GPR signal in an embodiment of the present invention, (a) is dry sand, (b) is dry loam, (c) is limestone, and (d) is a mixed medium;
[0049] Fig. 9 Schematic diagram of identifying characteristic signals using the wavelet transform modulus maximum method in an embodiment of the present invention, (a) is dry sand, (b) is dry loam, (c) is limestone, and (d) is a mixed medium;
[0050] Fig.10 Schematic diagram of dielectric constant distribution of a single-channel signal of a filling medium in an embodiment of the present invention;
[0051] Fig.11 is a histogram of residual modulus values of a dielectric constant fitting curve in an embodiment of the present invention. DETAILED DESCRIPTION
[0052] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.
[0053] Embodiment 1: A forward analysis method for dielectric properties of a mixed medium ground penetrating radar comprises the following steps:
[0054] Step 1: Divide the coarse grid model of the mixed medium established based on FDTD into several fine grid models that meet the boundary conditions, and assign corresponding dielectric parameters to the Yee cells in each fine grid model;
[0055] Step 2: Use the random shuffle algorithm to randomly shuffle all Yee cells with equal probability, and reassemble the shuffled Yee cell array into a coarse grid model in the form of a differential grid;
[0056] Step 3: Program and run the FDTD algorithm program on the MATLAB language platform to generate a geoelectric model of the mixed medium and obtain the GPR image under the pulse electromagnetic wave excitation state;
[0057] Step 4: Use image entropy theory to quantitatively describe the entropy value of the mixed medium GPR image, and use waveform analysis and wavelet transform modulus maximum method to analyze the GPR single-channel signal of the mixed medium.
[0058] Specifically, the fine grid model is as follows Figure 2As shown in (a), the electric field is located vertically inward from the center of the grid, and the magnetic field is located on the edge of the grid. The arrow indicates the magnetic field value of the fine grid. The circular particles represent media with different dielectric properties and are placed in the fine grid. Based on the basic principle of the shuffle algorithm, media with different dielectric properties are randomly mixed and spread over the entire fine grid model. The specific implementation steps are as follows: First, determine the number and dielectric parameters of the mixed media in the forward model. Taking the three mixed media {x}, {y}, and {z} as examples, the corresponding conductivity, magnetic permeability, and dielectric constant are {σ x , μ x , ε x}、{σ y , μ y , ε y}、{σ z , μ z , ε z}; Secondly, according to the restrictive conditions of the spatial step and time step of the grid model, the shapes and sizes of the three types of media are set. Here, the medium grid model is set to a circular particle shape, and air is filled between each particle. The particle size meets the requirements of the minimum spatial step of the fine grid model; then, according to the basic principle of the shuffle algorithm, the circular particles of the three mixed media are placed in a rectangular box with a perfectly matched layer boundary absorption condition, and the random shuffle algorithm is used to randomly sort the particle model, so as to obtain the scrambled matrix coordinates X = {σ in , μ in , ε in} and Y={σ in , μ in , ε in}, where i represents the type of medium and n represents the particle number; finally, the matrix position information generated by the rearrangement of all particles is used to generate a plaintext image sequence A row by row to obtain the scrambled matrix R, and the FDTD method is used to perform forward simulation imaging of the mixed medium using ground penetrating radar.
[0059] After the above steps, the forward simulation of the ground penetrating radar image of the dispersive mixed medium can be realized. The random shuffle algorithm is used to scramble the medium particles to obtain a three-dimensional two-dimensional fine grid distribution, such as Figure 2 (b) After optimization by the random shuffle algorithm, the generation probability of the filling medium placed in the rectangular box remains consistent, which can be used as the original model image of the ground penetrating radar for mixed medium forward simulation.
[0060] The propagation of electromagnetic waves is a quasi-spherical wave. The field value of the electromagnetic wave at the point near the source is larger than that at the point far from the source, and the field value of the electromagnetic wave decreases with the increase of distance. In the GPR simulation of homogeneous media, the amplitude of the electromagnetic wave decays exponentially with the propagation distance. For non-uniform media such as mixed media, the amplitude of the electromagnetic wave decays nonlinearly with the propagation distance, and the energy of the electromagnetic wave also exhibits nonlinear attenuation. This makes it impossible for the conventional two-dimensional Yee coarse grid model to express the electromagnetic response characteristics of the mixed medium. Therefore, constructing a fine grid model is the best way to construct a mixed medium model. The wave field component distribution of the two-dimensional Yee grid, such as Figure 1 As shown in the figure, the spatial layout of the fine grid algorithm is consistent with the conventional FDTD algorithm. The electric field is located vertically inward from the center of the grid, and the magnetic field is located on the edges of the grid. The large arrows represent the magnetic field values of the coarse grid, and the small arrows represent the magnetic field values of the fine grid. The ratio between the fine grid and the coarse grid follows the minimum grid restriction condition.
[0061] The FDTD method uses the solution of the difference equation group to replace the solution of the original electromagnetic field partial differential equation group, but the solution of the discretized difference equation group must meet the conditions of convergence and stability. The conditions of convergence and stability actually have certain restrictions on the space step and time step.
[0062] Specifically, the necessary conditions for the spatial step length are:
[0063]
[0064] The stability condition for the time step is:
[0065]
[0066] Among them, Δx and Δy are the spatial steps in the x and y directions respectively; Δt is the time step; c is the propagation speed of electromagnetic waves in the air, 0.3m / ns; λ represents the transmission wavelength of the electromagnetic wave.
[0067] In addition, the Fisher–Yates random shuffle algorithm is an algorithm for generating a random arrangement of a finite set (randomly sorting an array). The random arrangement generated by this algorithm is of equal probability. Its principle is similar to putting all array elements into the same deck of playing cards and generating a random arrangement sequence, and then determining the next element by continuously extracting elements from a deck of playing cards until there are no elements. The specific description is as follows: It is known that there are n cards, marked as a [n] , the serial numbers are 0, 1, 2, ..., n-1, start random shuffling, and randomly draw the sth card a from 1 to n-2 cards [s] , randomly match this card with a [0] or a [n-1]Exchange positions, and play card A [s] Put in an empty card sequence b [j] (j=0,1,2,…,n-1), let b [0] =a [s] , continuously extract from the original card until no more cards can be extracted, generating a new sequence b [j] , then b [j] is a random sequence generated by the [i] shuffling algorithm. For example, in the sequence A = {1,2,3,4,5,6,7}, a card number is randomly drawn 7 times each time (s = 3,4,1,3,1,1).
[0068] Step 1: S 0 =3, a 3 Random with a 0 or a 6 Exchange positions, here with a 0 If we swap positions, then A 0 ={2,3,1,5,6,7}, b 0 ={4};
[0069] Step 2: S 1 =4, a 4 Random with a 0 or a 5 Exchange positions, here with a 0 If we swap positions, then A 1 ={3,1,5,2,7}, b 1 ={4,6};
[0070] Step 3: S 3 =1, a 1 Random with a 0 or a 4 Exchange positions, here with a 0 If we swap positions, then A 2 ={3,5,2,7}, b 2 ={4,6,1};
[0071] Step 4: S 4 =3, a 3 Random with a 0 or a 3 Exchange positions, here with a 0 If we swap positions, then A 3 ={5,2,3}, b 3 ={4,6,1,7};
[0072] Step 5: S 5 =1, a 1 Random with a 0 or a2 Exchange positions, here with a 0 If we swap positions, then A 4 ={5,3}, b 4 ={4,6,1,7,2};
[0073] Step 6: S 6 =1, a 1 Random with a 0 or a 1 Exchange positions, here with a 0 If we swap positions, then A 5 ={5}, b 5 ={4,6,1,7,2,3}.
[0074] After the above steps, the remaining element is A 5 ={3}, then the generated sequence is b [j] ={4,6,1,7,2,3,5}, which is a random scrambled sequence obtained by the shuffling algorithm, and the probability of drawing a card each time is 1 / 7.
[0075] Embodiment 2: A forward analysis system for dielectric properties of a mixed medium ground penetrating radar includes a model segmentation module, a random scrambling module, an image acquisition module and a property analysis module.
[0076] Among them, the model partitioning module is used to partition the FDTD coarse grid model into several fine grid models that meet the boundary conditions, and assign corresponding dielectric parameters to the Yee cells in each fine grid model; the random scrambling module is used to use the random shuffling algorithm to randomly scramble all Yee cells with equal probability, and recombine the scrambled Yee cell array into a coarse grid model in the form of a differential grid; the image acquisition module is used to program and run the FDTD algorithm program on the MATLAB language platform to generate a geoelectric model of the mixed medium, and obtain the GPR image under the pulse electromagnetic wave excitation state; the characteristic analysis module is used to use the image entropy theory to quantitatively describe the GPR image entropy value of the mixed medium, and use the waveform analysis method and the wavelet transform modulus maximum method to analyze the GPR single-channel signal of the mixed medium.
[0077] Example 3
[0078] 1. Forward Modeling
[0079] Ground penetrating radar mainly consists of three parts: a control unit, a transmitter, and a receiver. The control unit transmits high-frequency electromagnetic waves to the detection target through the transmitting antenna. Due to the difference in dielectric properties between the target and the background medium, the electromagnetic waves are reflected and refracted when they propagate to the target, and follow the law of electromagnetic wave reflection and refraction. Finally, the electromagnetic wave signal is transmitted to the receiving antenna on the ground. The control unit converts the electromagnetic signal into a two-dimensional digital signal, and forms a ground penetrating radar image to be displayed on the control unit's display. The other part of the electromagnetic wave penetrates the target and diffuses underground. Therefore, the detection principle of ground penetrating radar is to transmit electromagnetic waves to the stratum along the survey line direction through a moving antenna, and then receive the electromagnetic wave signal reflected back from the target. The main electromagnetic display parameters are: two-way travel time, intensity and energy. The principle diagram of ground penetrating radar detection imaging is as follows: Figure 3 shown.
[0080] 2. Geological radar geoelectric model
[0081] Assume that the two-dimensional plane is a semi-infinite space continuous homogeneous medium, the electromagnetic wave reflection and refraction are all carried out in the plane, the background medium material is concrete, and dry sand, dry loam and limestone are filling media. In order to study the reflection characteristics of the mixed medium fine grid model on the GPR forward simulation signal, the following is designed: Figure 4 Two sets of geoelectric models for single medium and mixed medium are shown.
[0082] The geoelectric parameters of GPR are set as follows: ① Area range: 2.0m×0.4m, air layer thickness is 0.01m, the lower left corner is the coordinate origin, the horizontal axis is the horizontal distance of the concrete model, and the vertical axis is the detection depth. 1 ,y 2 ,y 3 ) length is 0.2m, and the heights are 0.1, 0.15, and 0.2m respectively; ② Medium type: concrete, dry sand, dry loam, and limestone, the relative dielectric constants are designed to be 6.0, 4.0, 8.0, and 12.0 respectively, and the corresponding dielectric parameters are shown in Table 1; ③ Number of simulations: four times in total, the first three times are three single media of dry sand, dry loam, and limestone, and the fourth time is a mixture of the three single media, all filled in rectangular boxes of different heights.
[0083] Table 1 Dielectric properties of media
[0084]
[0085] The mixed medium is composed of a circular model with a radius of 2.5 mm, a rectangular box y 1 ,y 2 ,y 3The number of particle models in the model is 800, 1200, and 1600 respectively. The three media are randomly mixed, and then the general FDTD modeling method is used to obtain the geoelectric model as follows Figure 4 As shown. Figure 4 (a) with Figure 4 (b) It can be seen that y 1 ,y 2 ,y 3 In the rectangular box, the grid model of the single medium shows that the internal medium is evenly distributed, while the fine grid model of the mixed medium shows the effect of random distribution of the internal medium, indicating that fine grid modeling is feasible.
[0086] 3. Ground Penetrating Radar Signal Processing
[0087] According to the relationship between the GPR antenna frequency, detection depth and resolution, the antenna center frequency is set to 1.6 GHz, the boundary absorption condition is a perfectly matched layer, the excitation source is Ricker wavelet, the spatial step and sampling step of the grid are both 0.0025 m, the number of sampling channels is 800, and the time window is 10 ns. The computing environment is an Intel(R) Core(TM) i7-4790K CPU @ 4.00 GHz, 4 GHz + 24 GB memory physical address extension, and a Lenovo desktop computer with Windows 7 operating system. The FDTD method is used to simulate the GPR characteristic response signals of three media, namely dry sand, dry loam and limestone, and the original GPR images are obtained, such as Figure 5 The original GPR images of the fine grid model of the mixed medium are shown in a, 5b, and 5c. Figure 5 d. Figure 5 As can be seen from a to 5c, the dielectric dispersion causes electromagnetic wave reflection and diffraction to occur at the lower interface of the model box, and the larger the dielectric constant of the medium, the wider the radiation range and the stronger the energy. Figure 5 In d, electromagnetic wave reflection and diffraction phenomena also occur inside and around the model box filled with mixed media due to medium dispersion, and the scattering energy and reflection amplitude are stronger.
[0088] The original GPR image is disturbed by diffraction waves and cannot clearly reflect the signal characteristics and actual position of the target. Therefore, it is necessary to perform signal preprocessing on the original radar image, which mainly includes static correction removal, gain, bandpass filtering, offset and other processes. After signal preprocessing, the GPR image of a single medium (such as Figure 5 e, 5f, 5g) and mixed medium ground penetrating radar images (as shown in Figure 5h), after signal preprocessing, the multiple wave reflections of the radar signal are suppressed, the curvature of the diffraction hyperbola converges, and the reflected energy is more concentrated, which can truly reflect the characteristic response signal of the ground penetrating radar of the filling medium. Especially for mixed media, the mixed medium particles in the rectangular box can better show the state of mixing between the media, which shows that the FRS-FDTD method can successfully simulate the ground penetrating radar response images of various mixed media.
[0089] 4. Image Entropy Processing
[0090] Entropy refers to the degree of disorder of a system, while image entropy is a characteristic statistical form that reflects the average amount of information in an image and also indicates the degree of disorder of information distribution in an image. The larger the entropy, the higher the degree of disorder, and vice versa. In the process of scrambling mixed media using the random shuffling algorithm, the local entropy of the image is defined according to the grayscale distribution of each pixel in the ground penetrating radar image. It can reflect the degree of focus of the image information and determine the degree of disorder of the fine grid model medium particle distribution. The specific operation is to extract the grayscale image of each filling medium test box, such as Figure 6 As shown, the image entropy calculation program is then written to obtain the entropy value of each grayscale image and statistically calculate it. Figure 7 middle.
[0091] from Figure 6 Comparison of the grayscale images of the various filling media shows that irregular wrinkles appear on the mixed medium image, indicating that the reflection and refraction of electromagnetic wave signals in the medium are active, while the images of other filling media are relatively pure and do not show obvious characteristic information. Figure 7 It can be seen that the image entropy value of concrete medium is the lowest, while the image entropy values of three filling media such as dry sand, dry loam and limestone are not much different, and the image entropy value decreases with the increase of filling medium height, indicating that the electromagnetic wave is reflected and refracted more frequently in the thinner strata, which easily leads to a more disordered state of the internal system. The mixed medium fine grid model constructed by FRS-FDTD method has the highest ground penetrating radar image entropy value, indicating that the electromagnetic response characteristics of the fine grid model are operating normally, the electromagnetic wave reflection and refraction between mixed media follow the laws of electromagnetic propagation, and the distribution between each medium tends to be disordered and has a high degree of random distribution, which can further analyze the dielectric properties of the GPR single-channel signal.
[0092] 5. Waveform Analysis
[0093] The waveform of electromagnetic waves can reflect some dielectric properties of the medium. By analyzing the time-amplitude diagram of the GPR signal, the dielectric parameters of the medium can be inferred. In order to explore the dielectric properties of ground penetrating radar with different dielectric fillings, a single channel signal at different longitudinal heights of the rectangular box was extracted, such as Figure 8 The time-width diagram shown. Figure 8(a~c) show that the direct wave signal of the air layer has been eliminated. At time point t 0 It is represented as the signal characteristic point when the electromagnetic wave passes through the upper interface of the rectangular box, and t 1 ,t 2 ,t 3 They represent the signal feature points when the electromagnetic wave passes through the experimental box at a longitudinal height of 10 cm, 15 cm, and 20 cm, respectively. The relevant data of the ground penetrating radar statistics are shown in Table 2. As can be seen from Table 2, the average dielectric constants of dry sand, dry loam, and limestone at heights of 10 cm, 15 cm, and 20 cm are 3.88, 7.99, and 12.01, respectively, and the relative errors with the design values of the dielectric constants are 3.00%, 0.125%, and -0.08%, respectively. The relative error values show a decreasing trend, indicating that within the detection depth range of the antenna center frequency, if the depth of the detected target is greater than the wavelength of the electromagnetic wave, the higher the detection accuracy.
[0094] Table 2 Radar signal recognition data
[0095]
[0096] exist Figure 8 In (d), it is not possible to clearly identify the feature points where the electromagnetic wave passes through the upper interface of the rectangular box, nor is it possible to pick up the feature points where the electromagnetic wave propagates to the lower interface of the rectangular box. The main reason for this is that the random distribution of the mixed medium particles and the air gaps between the medium particles lead to multiple reflections and refractions during the propagation of the electromagnetic wave, making the feature point signals submerged in the diffraction wave signals. Therefore, an algorithm is needed to separate the interference of the clutter signal and accurately identify the singular points of the GPR signal.
[0097] VI. Wavelet Transform Modulus Maximum Analysis
[0098] There is a close relationship between the modulus maximum method of wavelet transform and the singular points of non-stationary signals. The singular points and irregular mutation parts of non-stationary signals often carry relatively important information, which is one of the important characteristics of the signal. For the non-destructive detection signal of ground penetrating radar, when the electromagnetic wave propagates in two or more media, a signal mutation phenomenon will occur at the critical surface of the heterogeneous medium, which is reflected as a singular point on the one-dimensional signal of the signal mutation surface, that is, the waveform peak point on the single-channel signal of the ground penetrating radar. The signal of the ground penetrating radar detecting mixed media is often a non-stationary signal, which is prone to overlapping interference of reflected waves and diffracted waves, which will have a serious impact on the identification of the peak point on the single-channel signal. Therefore, the use of the modulus maximum method of wavelet transform can reduce the interference of surrounding clutter and accurately identify the singular points of the GPR single-channel signal.
[0099] The single-channel signals of the ground penetrating radar of the filling medium at different heights were extracted respectively, and the wavelet transform modulus maximum method program was written on the MATLAB language platform. The recognition results are as follows: Fig. 9As shown. Fig. 9 The time point t when the electromagnetic wave propagates to the upper and lower interfaces of the rectangular box can be clearly identified. 0 With t 1 , the dielectric constant can be calculated when the height of the rectangular box is known. After calculation, the average dielectric constants of dry sand, dry loam, limestone and mixed media in different height directions are 3.83, 7.88, 12.01 and 6.87 respectively. Among them, the relative errors between the calculated value and the designed value of the dielectric constant of a single medium are 4.25%, 1.5% and -0.83% respectively, and the relative error rate is controlled within 5%. Within the maximum depth range that can be detected by the radar antenna frequency, as the transmission distance of the electromagnetic wave in the medium increases, the relative error rate of identification also decreases. Since there is no corresponding design value for comparison, the relative error rate of the mixed medium cannot be calculated directly. Therefore, we propose to expand the number of samples to find the optimal solution for the dielectric constant. The specific operation is to use the wavelet transform modulus maximum method to identify a total of 240 single-channel signals in the longitudinal direction of the four types of media, and calculate the data statistics. Fig.10 middle.
[0100] from Fig.10 It can be seen that the dielectric constant distribution of dry sand, dry loam and limestone is relatively uniform, while the dielectric constant distribution of mixed media fluctuates greatly. Specifically, the dielectric constant variation increases with decreasing height. This also shows that the reflection and refraction phenomenon of electromagnetic waves in a small distance range is more obvious, and is similar to the variation law of the dielectric constant of a single medium at different heights. Further, the dielectric constant scatter plots of the four types of media are linearly fitted and analyzed, and the residual modulus bar graph of the dielectric constant of each single signal is obtained, as shown in Fig.11 As shown in . The smaller the residual modulus value, the higher the degree of fit and the tighter the distribution. Conversely, the lower the degree of fit and the more discrete the distribution. Fig.11 It is found that the residual value changes the most at the critical point of height change, which is due to the incomplete energy convergence caused by dielectric dispersion at the boundary of rectangular boxes of different heights. Moreover, the residual modulus of the dielectric constant of the mixed medium is more than twice that of the single medium, indicating that the discrete degree of the dielectric constant distribution of the mixed medium is much greater than that of the single medium, which is consistent with the dielectric constant distribution characteristics of the actual mixed medium.
[0101] VII. Conclusion Analysis
[0102] By comparing and analyzing the GPR images and signals of single media and mixed media, it is found that the entropy value of the GPR image of the fine grid model is higher, indicating that the propagation path and reflection law of the radar electromagnetic wave in the mixed medium are more complicated, which is more in line with the ground penetrating radar image characteristics of the mixed medium in the actual stratum. The waveform analysis method can identify the location of the singular point of the GPR single-channel signal, mainly relying on the characteristics of the dielectric constant gradient change between different media. The dielectric constant of a single dielectric body is uniformly distributed, and only changes with the dielectric constant of the background medium in a gradient, making it easier to find the location of the singular point, such as Figure 8 (a) to (c) are shown. However, the dielectric constant of the mixed medium is non-uniformly distributed. Not only does the dielectric constant at the interface with the background medium change in gradient, but also the dielectric constant gradient changes between different media inside. This makes it impossible to directly pick up singular points by waveform analysis, such as Figure 8 As shown in (d), this is also one of the reasons why the recognition accuracy is not high in actual GPR detection.
[0103] The wavelet transform modulus maximum method can effectively reduce the electromagnetic interference between different electrical media, effectively identify the characteristic points at the boundary between the mixed medium and the background medium, and obtain the two-way travel time of the electromagnetic wave based on the difference between the horizontal coordinates of the two characteristic points, thereby inversely calculating the relative dielectric constant of the single-channel signal. Fig. 9 In (d), the modulus maximum curve of the mixed medium can clearly observe two peak points, which are also characteristic points. Fig. 9 (a)~(c), although there are still some clutter signals on the curve, it does not affect the recognition of feature points. Fig.10 The dielectric constant distribution of four different types of media is shown. The dielectric constant scatter plot of the mixed media shows an uneven distribution, which indicates that the dielectric properties of some media are retained after the multi-media are mixed, but the dielectric constants of the media do not show an obvious relationship. Fig.11 In the figure, the residual modulus of the dielectric constant between adjacent single-channel signals fluctuates greatly, and there is no obvious regularity in the distribution of dielectric constants between different media, which indirectly illustrates the effectiveness of the random shuffle algorithm.
[0104] The present invention constructs a mixed medium fine grid model for the first time based on the FRS-FDTD method, and successfully realizes the GPR simulation of the mixed medium. Compared with the general FDTD grid model, the fine grid model of the mixed medium has the characteristics of electrical compatibility and random distribution. Through FDTD simulation, the GPR two-dimensional time-distance profile and electromagnetic response signal of the mixed medium are obtained, and it is found that the electromagnetic wave propagation path and electrical parameter distribution of the mixed medium are more complex than those of the single medium. Specifically, the entropy value of the ground penetrating radar image of the mixed medium is higher than that of the single medium, indicating that the internal dielectric parameters are discretely distributed after the multi-media are mixed, which is consistent with the distribution characteristics of the dielectric constant of the actual mixed medium, proving the effectiveness of the fine grid model. The single-channel signal of the mixed medium cannot effectively extract the signal feature points, which is consistent with the characteristics of the electromagnetic wave waveform overlap inside the mixed medium. The wavelet transform modulus maximum method is used to identify the characteristic signals of the single medium and the mixed medium respectively. By comparison, it is found that the dielectric constant of the single medium is evenly distributed, while the dielectric constant of the mixed medium is non-uniformly distributed. This phenomenon also proves again that the FDTD simulation image is consistent with the characteristics of the actual mixed medium. The residual value of the dielectric constant fitting of the mixed medium is twice that of the single medium, indicating that the dielectric constant distribution of the mixed medium is more discrete, which illustrates the effectiveness of random scrambling of the random shuffling algorithm in fine grid model construction.
[0105] In summary, we proposed a new method to construct a mixed medium model using the FDTD algorithm, and verified the effectiveness and stability of the method from the aspects of image entropy, electromagnetic wave waveform and dielectric constant of the mixed medium. This method can provide technical guidance for mastering the GPR electromagnetic response characteristics and dielectric properties of mixed media. Considering that GPR detection of mixed media in actual engineering is affected by multiple factors, it is necessary to improve the fine grid model of multi-media mixture in the future, increase the uncertainty influencing factor, and reduce the influence of medium dispersion in the fine grid model.
[0106] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.
[0107] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0108] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0109] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.
[0110] The above specific implementation methods further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above are only specific implementation methods of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. The forward analysis method of dielectric properties of mixed medium ground penetrating radar is characterized by: The following steps are involved: The coarse grid model of mixed medium established based on FDTD is divided into several fine grid models that meet the boundary conditions, and the corresponding dielectric parameters are assigned to the Yee cells in each fine grid model. The random shuffle algorithm is used to randomly shuffle all Yee cells with equal probability, and the shuffled Yee cell array is reassembled into a coarse grid model in the form of a differential grid; Run the FDTD algorithm program to generate the geoelectric model of the mixed medium and obtain the GPR image under the pulse electromagnetic wave excitation state; Image entropy theory is used to quantitatively describe the entropy value of GPR images in mixed media, and waveform analysis method and wavelet transform modulus maximum method are used to analyze the GPR single-channel signal of mixed media.
2. The forward analysis method for dielectric properties of mixed medium ground penetrating radar according to claim 1 is characterized in that: The Yee cell assignment process in the fine grid model is specifically as follows: Determine the number of mixed media and dielectric parameters in the fine grid model; Assign corresponding dielectric parameters to each mixed medium.
3. The forward analysis method for dielectric properties of mixed medium ground penetrating radar according to claim 1 is characterized in that: The dielectric parameters include electrical conductivity, magnetic permeability and dielectric constant.
4. The forward analysis method for dielectric properties of mixed medium ground penetrating radar according to claim 1 is characterized in that: The subdivision process of the fine grid model is specifically as follows: Under the condition that the solution of the finite difference time domain equation system has stable convergence, the restrictive conditions of the space step and time step of the grid model are defined; The shape, size and dielectric properties of the mixed medium are determined according to restrictive conditions to achieve the minimum division of the fine grid model.
5. The forward analysis method for dielectric properties of mixed medium ground penetrating radar according to claim 4 is characterized in that: The calculation formula of the restrictive condition is specifically: The necessary condition for the spatial step size is: The stability condition for the time step is: Among them, Δx and Δy are the spatial steps in the x and y directions respectively; Δt is the time step; c is the propagation speed of electromagnetic waves in the air, 0.3m / ns; λ represents the transmission wavelength of the electromagnetic wave.
6. The forward analysis method for dielectric properties of mixed medium ground penetrating radar according to claim 1 is characterized in that: The process of Yee cells being randomly shuffled with equal probability is specifically as follows: The specific shapes of various mixed media are placed into a rectangular box with a perfect matching layer as the boundary absorption condition; The random shuffle algorithm is used to randomly sort the shape model to obtain the scrambled matrix coordinates of the scrambled shape model.
7. The forward analysis method for dielectric properties of mixed medium ground penetrating radar according to claim 1 is characterized in that: The acquisition process of the GPR image is specifically as follows: The matrix position information generated by rearranging all shapes is used to generate the plaintext image sequence A row by row, and the scrambled matrix R is obtained; The FDTD method is combined to carry out forward simulation imaging of ground penetrating radar on the mixed medium to obtain the GPR image.
8. The forward analysis system of dielectric properties of mixed medium ground penetrating radar is characterized by: include: Model segmentation module, used to segment the FDTD coarse grid model into several fine grid models that meet the boundary conditions, and assign corresponding dielectric parameters to the Yee cells in each fine grid model; The random scrambling module is used to use the random shuffling algorithm to randomly scramble all Yee cells with equal probability, and reassemble the scrambled Yee cell array into a coarse grid model in the form of a differential grid; An image acquisition module is used to run the FDTD algorithm program to generate a geoelectric model of the mixed medium and obtain the GPR image under the pulse electromagnetic wave excitation state; The characteristic analysis module is used to quantitatively describe the entropy value of the mixed medium GPR image using the image entropy theory, and to analyze the GPR single-channel signal of the mixed medium using the waveform analysis method and the wavelet transform modulus maximum method.
9. A computer terminal comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the forward analysis method for dielectric properties of a mixed medium ground penetrating radar as described in any one of claims 1 to 7 is implemented.
10. A computer readable medium having a computer program stored thereon, characterized in that: The computer program is executed by a processor to implement the forward analysis method for dielectric properties of mixed medium ground penetrating radar as described in any one of claims 1 to 7.
