Ground penetrating radar ringing noise suppression method in non-uniform environment

Abstract

The invention belongs to the technical field of ground penetrating radars, and particularly relates to a method for suppressing ringing noise of a ground penetrating radar in a non-uniform environment. The method specifically comprises the following steps: step 1, establishing a non-uniform environment model, and simulating ringing noise in the environment model; 2, constructing a background ringing noise matrix for suppressing the ringing noise, and obtaining echo data B after the ringing noise is removed; 3, for the echo data matrix B obtained in the step 2, reducing the intensity of non-uniform clutters by using a singular value decomposition method to obtain an echo data matrix C; 4, aiming at the echo data matrix C, inhibiting non-uniform clutters by using symmetric filtering to obtain an echo data matrix S; and 5, taking the peak signal-to-noise ratio between the echo data before processing and the echo data after processing in the steps 1-4 as a target function, selecting parameters through a Bayesian optimization algorithm, obtaining a parameter value of an echo data matrix when the peak signal-to-noise ratio is maximum, and finally realizing ground penetrating radar ringing noise suppression.

Description

Technical Field

[0001] The present invention belongs to the technical field of ground penetrating radar, and particularly relates to a method for suppressing ringing noise of ground penetrating radar in a non-uniform environment. Background Art

[0002] Ground Penetrating Radar (GPR) is a non-destructive detection technology that has developed rapidly in recent years. Its principle is to transmit high-frequency electromagnetic waves into the ground through a transmitting antenna. When the electromagnetic waves propagate to a place with obvious differences in underground physical properties, reflection and refraction will occur. The receiving antenna receives these reflected echoes, and after signal processing, an echo signal image is formed, thereby realizing the detection of underground targets. Because of the advantages of ground penetrating radar such as a wide range of target material types, high resolution, real-time imaging, and non-destructive detection, it has been widely used in fields such as highway quality inspection, municipal pipeline detection, geological and hydrological monitoring, building damage detection, and exploration. However, due to the physical structure and electrical characteristics of the transmitting antenna itself, a ringing phenomenon will occur. When the antenna transmits a pulse signal, not all of the energy in the antenna is released. A part of the energy will flow back and forth between the tip of the transmitting antenna and the input feeder. Each flow will generate a secondary pulse, resulting in the antenna sending multiple gradually attenuating pulses, forming a ringing phenomenon. Due to the impedance mismatch between the antenna and the ground, the multiple gradually attenuating pulses transmitted form reflected echoes on the ground, which are manifested as horizontal stripes covering the entire range of the echo image in the echo image. Generally, the ringing noise masks the hyperbolic characteristics of the target echo, affecting the quality of the ground penetrating radar echo image. The ground penetrating radar echo signal is also affected by non-uniform clutter generated by changes in soil conditions and surface roughness, making it more difficult to detect the target echo signal.

[0003] To suppress the non-uniform clutter and ringing noise in ground penetrating radar, researchers have proposed a variety of traditional methods, mainly including mean cancellation method, median cancellation method, singular value decomposition method (SVD), discrete wavelet transform method (DWT), wavenumber domain filtering method, and deep learning-based method. The mean cancellation method and median cancellation method have simple principles and are easy to implement. However, when used for suppressing ringing noise, they are prone to generating new horizontal stripes at the vertices of target echoes and have poor suppression effects on non-uniform clutter. The singular value decomposition method has a poor suppression effect on ringing noise, and parameter selection depends on manual experience. The discrete wavelet transform method has the problem of difficult selection of wavelet basis functions, and different wavelet basis functions will have a significant impact on the processing results. The wavenumber domain filtering method can perform targeted filtering according to the frequency of ringing noise. However, for non-uniform clutter, a fixed wavenumber domain filter cannot effectively suppress it, and more complex adaptive filtering methods need to be used. The deep learning-based method can adaptively achieve the best method for noise suppression through a large amount of training data, and its noise suppression effect is good. However, it requires a large amount of data for training, a large amount of computing resources and time, and the generalization performance of the model is poor, resulting in poor noise suppression effects in new scenarios. Summary of the Invention

[0004] To solve the above problems, the present invention proposes a method for suppressing ringing noise of ground penetrating radar in a non-uniform environment. Compared with the above traditional methods, the present invention solves the problems of difficult model and parameter selection and poor generalization performance of the method in different scenarios, and has good suppression effects on non-uniform clutter and ringing noise.

[0005] The technical solution of the present invention is implemented as follows:

[0006] A method for suppressing ringing noise of ground penetrating radar in a non-uniform environment, the specific process is as follows:

[0007] Step 1: Establish a non-uniform environment model and simulate ringing noise in the environment model;

[0008] Step 2: Construct a background ringing noise matrix to suppress the ringing noise and obtain the echo data B after removing the ringing noise;

[0009] Step 3: For the echo data matrix B obtained in Step 2, use the singular value decomposition method to reduce the intensity of non-uniform clutter and obtain the echo data matrix C;

[0010] Step 4: For the echo data matrix C, use symmetric filtering to suppress non-uniform clutter and obtain the echo data matrix S;

[0011] Step 5: Taking the peak signal-to-noise ratio between the pre-processed echo data and the echo data after being processed by Steps 1-4 as the objective function, select parameters through the Bayesian optimization algorithm to obtain the parameter values of the echo data matrix at the maximum peak signal-to-noise ratio, and finally realize the ringing noise suppression of the ground penetrating radar.

[0012] Optionally, the specific method for simulating the ringing noise in the present invention is: in a non-uniform environment model, make the receiving antenna and the transmitting antenna at the same horizontal position, place an iron plate at a set distance from the transmitting antenna to reflect the pulse signal emitted by the transmitting antenna, and the signal received by the receiving antenna simulates the ringing noise.

[0013] Optionally, the specific process of Step 2 in the present invention is as follows:

[0014] First, take a horizontal window with a length of l, traverse the echo data matrix by sliding the window, and trim and weight the echo data within the window to preliminarily remove the target echo signal and obtain the background ringing noise matrix;

[0015] Secondly, perform iterative optimization on the generated background ringing noise matrix until all target echo signals are removed, leaving only the background ringing noise matrix;

[0016] Finally, subtract the background ringing noise matrix from the original signal echo data to obtain the echo data after removing the ringing noise.

[0017] Optionally, the pruning in the present invention is: removing the smallest and largest partial values within the window according to the pruning ratio α, and taking the average of the remaining values within the window to obtain the average value A of the window pruned by the ratio α α ;

[0018] The weighting is: setting a custom parameter s, setting the normalization weight according to the difference between the remaining values within the window and A α and then performing weighted summation to obtain the value A of the background ringing noise at the center of the window rc .

[0019] Optionally, the weight of the weighting in the present invention is:

[0020]

[0021] where s is a custom parameter, and d i is the difference between the amplitude a i of the i-th element in the window and A α ;

[0022] Then normalize the weight W i to obtain the normalized weight

[0023] Optionally, in step 3 of the present invention, reducing the intensity of non-uniform clutter by using the singular value decomposition method is as follows: for the strip-shaped non-uniform clutter generated by the rough surface, select the previous or the first few larger singular values to remove; for the non-uniform clutter generated by the underground sand and gravel, select the smaller singular values or the singular values close to 0 to remove.

[0024] Optionally, the specific process of step 4 of the present invention is as follows: traverse all elements in matrix C with a sliding window to obtain the position J with the strongest symmetry in each row of matrix C m , calculate the position J with the strongest symmetry in each row m symmetric weight matrix α(i), calculate the symmetric weight matrix β[i,j] at each element of data matrix C; set the influence factors γ and μ of the two weights, calculate the synthetic symmetric filtering weighted matrix w[i,j], and use the matrix w[i,j] to perform symmetric filtering on matrix C to suppress non-uniform clutter.

[0025] Optionally, the specific process of step 4 of the present invention is as follows:

[0026] Calculate the symmetric weight matrix α(i) at J m in each row, and perform normalization processing on it;

[0027]

[0028] where J 0 (i) represents the column number at the position with the strongest symmetry in the i-th row;

[0029] Calculate the symmetric weight matrix β[i,j] at each element of data matrix C, and perform normalization processing;

[0030]

[0031] Optionally, in step 4 of the present invention, calculate the synthetic symmetric filtering weighted matrix w[i,j], and use the matrix w[i,j] to perform symmetric filtering on matrix C to suppress non-uniform clutter; the specific process is as follows:

[0032] Calculate the synthetic symmetric filtering weighted matrix w[i,j], and perform symmetric filtering to suppress non-uniform clutter, which is expressed as:

[0033] w[i,j] = e γ·α[i] ·e μ·β[i,j]

[0034] s[i,j] = c[i,j] · w[i,j]

[0035] where γ and μ are the influence factors of the symmetric weight matrices α[i] and β[i,j], and s[i,j] is the signal after symmetric filtering to suppress non-uniform clutter.

[0036] Optionally, in step 5 of the present invention, for the sliding window length l, pruning ratio α, weighting intensity s, and the symmetric weight matrix influence factors γ and μ in the symmetric filtering process, the parameters are selected by the Bayesian optimization algorithm.

[0037] Beneficial effects:

[0038] First, according to the characteristics of ringing noise, non-uniform clutter, and target echoes, the present invention sets a horizontal sliding window in the received signal data matrix, and performs pruning and weighting processing on the signals within the window. By continuously iteratively updating the signals within the window, the effect of suppressing non-uniform clutter and ringing noise is achieved.

[0039] Second, when suppressing non-uniform clutter, the present invention first uses the singular value decomposition method to reduce the intensity of non-uniform clutter, improving the effect of suppressing non-uniform clutter. For the selection of some key parameters, the present invention uses the Bayesian optimization algorithm for automatic selection, solving the problem of difficult selection of key parameters. Through simulation experiment verification, in a non-uniform environment model, the present invention has a good suppression effect on both non-uniform clutter and ringing noise. Description of the drawings

[0040] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0041] Figure 1 is the flow chart of the present invention;

[0042] Figure 2 is the non-uniform environment simulation model diagram established by the method of the present invention;

[0043] Figure 3 is the simulation result diagram of the method of the present invention in a non-uniform environment; Figure 3 (a) is the B-scan image obtained by gprMax simulation, Figure 3 (b) is the echo data after adding ringing noise, Figure 3 (c) is the added ringing noise;

[0044] Figure 4 is the ringing noise suppression result diagram of the method of the present invention; Figure 4 (a) is the background ringing noise constructed by the method in step 2 of the present invention, Figure 4 (b) is the result after suppressing the ringing noise;

[0045] Figure 5 is the non-uniform clutter suppression result diagram of the method of the present invention;Figure 5 (a) is the result after suppressing the ringing noise and then performing singular value decomposition. Figure 5 (b) is Figure 5 based on the singular value decomposition in (a) and performs symmetric filtering to suppress non-uniform clutter.

[0046] Specific implementation process

[0047] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0048] It should be noted that, without conflict, the following embodiments and the features in the embodiments may be combined with each other; and, based on the embodiments in the present disclosure, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present disclosure.

[0049] It should be noted that the following describes various aspects of embodiments within the scope of the appended claims. It should be apparent that the aspects described herein may be embodied in a wide variety of forms, and any specific structure and / or function described herein is illustrative only. Based on the present disclosure, those skilled in the art should understand that one aspect described herein may be implemented independently of any other aspect, and two or more of these aspects may be combined in various ways. For example, any number of aspects described herein may be used to implement the device and / or practice the method. In addition, this device and / or this method may be implemented using other structures and / or functions in addition to one or more of the aspects described herein.

[0050] An embodiment of the present application is a method for suppressing ringing noise of a ground penetrating radar in a non-uniform environment, as Figure 1 shown, the specific process is as follows:

[0051] Step 1: Establish a non-uniform environment model and simulate ringing noise in the environment model;

[0052] The non-uniform environment model in this step is constructed by the Peplinski soil model in the gprMax simulation software. This model can simulate the non-uniform environment in real soil, taking into account the influence of components such as sand, clay, and moisture in the soil on the overall dielectric constant of the soil. Different non-uniform environment models can be established by controlling the proportion of sand and clay and the range of volumetric water fraction.

[0053] Ringing noise is caused by the fact that when the antenna emits pulsed signals, not all the energy in the antenna is released. A part of the energy will flow back and forth between the tip of the transmitting antenna and the input feeder. Each flow generates a secondary pulse, resulting in the antenna sending multiple gradually attenuating pulses. Due to the impedance mismatch between the antenna and the ground, multiple gradually attenuating pulses are reflected on the ground to form multiple ground reflection waves, thus forming ringing noise. In this step, the gprMax simulation software is used to simulate the ringing noise. A steel plate is placed 5 cm away from the transmitting antenna to reflect the pulsed signals emitted by the transmitting antenna. The receiving antenna is at the same horizontal position as the transmitting antenna, with a 10 cm interval between them. The signals received by the receiving antenna are used to simulate the ringing noise.

[0054] Step 2: Construct a background ringing noise matrix to suppress the ringing noise and obtain the echo data after removing the ringing noise;

[0055] Ringing noise appears as horizontal bands with relatively large amplitudes in the radar echo signal image, almost completely masking the target echo signal and non-uniform clutter. First, according to the characteristics of the horizontal distribution of the ringing noise, the echo data matrix is traversed through a sliding window, and the echo data within the window is trimmed and weighted to preliminarily remove the target echo signal, obtaining a background ringing noise matrix; secondly, the generated background ringing noise matrix is iteratively optimized until all the target echo signals are removed, leaving only the background ringing noise matrix; finally, it is subtracted from the original signal echo data to obtain the echo data after removing the ringing noise. The detailed steps are as follows:

[0056] In the radar echo signal matrix, a horizontal window with a length of l is taken starting from the first row. The window length can be changed according to different radar echo data, and the number of elements within the window is denoted as L. The elements within the window are arranged in ascending order, and then the smallest and largest partial values within the window are removed according to the set ratio α. Since the amplitude change of the ringing noise within the window is very small, it can be considered that the extreme values removed are generated by the target echo signal and non-uniform clutter. The average value of the remaining values within the window is taken to obtain the average value of the window trimmed by the ratio α, denoted as A α It is expressed as:

[0057]

[0058] where, a i is the amplitude of the i-th element, and N = αL.

[0059] The obtained average value A α can be used as a reference value for the background ringing noise at the center of the window. However, due to the randomness of the setting of the α value, there may be some extreme values in the window that are not removed, thus affecting the average value. Therefore, A αThe true value equivalent to the background ringing noise. The solution is to set the normalization weights according to the differences between the remaining values within the window and A α and then perform weighted summation to obtain the value A rc of the background ringing noise at the center of the window, where r and c represent the row and column where the center position of the window is located. The weight W i of the i-th element within the window is expressed as:

[0060]

[0061] where s is a user-defined parameter representing the weighting intensity set according to the difference between a i and A α . d i is the amplitude a i of the i-th element in the window minus the value of A α , expressed as: Then, the weight W i is normalized to obtain the normalized weight and then is used to perform weighted summation on the amplitudes within the window to obtain the value A rc of the background ringing noise at the center of the window. The above process is expressed as:

[0062]

[0063] The horizontal window is slid along each row until the entire echo data matrix is traversed to initially obtain the background ringing noise matrix A. Then, the same process as above is performed on A. After iterating a certain number of times until all target echoes and non-uniform clutter are removed, a matrix containing only background ringing noise is obtained. Then, the original echo data matrix is subtracted from the background ringing noise matrix to achieve the effect of suppressing ringing noise.

[0064] Step 3: For the echo data matrix B obtained in Step 2, use the singular value decomposition method to reduce the intensity of non-uniform clutter and obtain the echo data matrix C;

[0065] The singular value decomposition method can decompose the above echo data matrix with removed ringing noise into the complementary subspaces of the target echo signal and the non-uniform clutter subspace using the properties of singular values, increasing the signal-to-noise ratio. Let the echo data matrix after removing ringing noise in Step 2 be B, and perform singular value decomposition on it as:

[0066] B = U × S × V T

[0067] where U and V are orthogonal matrices, S is the singular value matrix, S = diag(σ 1 , σ 2 ,..., σ r ), r = rank(B). BT The eigenvalue of B is λ i and is arranged in descending order. Let be called the singular value of matrix B.

[0068] Matrix B can be expressed through singular values as:

[0069]

[0070] B = B 1 + B 2 +... + B k

[0071] where B i is a matrix with the same dimension as B and is called the eigenimage of B.

[0072] In the present invention, according to the distribution characteristics of non-uniform clutter and the energy strength compared with the target echo, the singular values corresponding to the non-uniform clutter can be flexibly selected for removal, so as to reduce the energy of the non-uniform clutter and highlight the characteristics of the target echo signal. For the strip-shaped non-uniform clutter generated by a rough surface, since its clutter energy is larger than that of the target echo, the previous one or the previous several larger singular values should be selected for removal; for the non-uniform clutter generated by underground sand and gravel, since its clutter energy is smaller than that of the target echo, smaller singular values or singular values close to 0 should be selected for removal.

[0073] Step 4: For the echo data matrix C, use symmetric filtering to suppress non-uniform clutter and obtain the echo data matrix S;

[0074] In the ground penetrating radar echo image, the target echo has obvious hyperbolic characteristics, and there is a symmetry on both sides of the vertex of the hyperbola. According to the symmetry difference between the target echo and the non-uniform clutter, symmetric filtering can be used to suppress the non-uniform clutter. The detailed steps are as follows:

[0075] First, let the data matrix after singular value decomposition in step 3 be C, and each element in it is expressed as c[i, j]. Traverse all elements in matrix C with a sliding window of length K and width M. In the sliding window, by multiplying and summing the element values at symmetric positions and taking the maximum value, the position with the strongest symmetry in each row of matrix C is obtained, which is most likely the position of the target echo hyperbola. The above method is expressed as:

[0076]

[0077] where J m is the position with the strongest symmetry in each row of matrix C.

[0078] Secondly, calculate J for each row mThe symmetric weight matrix α(i) at [location] is expressed as:

[0079]

[0080] After obtaining α[i], it is normalized to eliminate the influence of the excessive weight at the vertex of the echo hyperbola on the entire symmetric filtering process.

[0081] Next, calculate the symmetric weight matrix β[i,j] at each element in the data matrix C, which is expressed as:

[0082]

[0083] Corresponding to α[i], β[i,j] also needs to be normalized. In addition, during the process of calculating the symmetric weight matrix, since there may be elements with a value of 0 in the data matrix C, it may cause the denominator of the symmetric weight matrix to be 0, resulting in calculation errors. For this, take the average value of the valid values in the two adjacent rows closest to the position with calculation errors to replace the value at the position with calculation errors.

[0084] Finally, calculate the composite symmetric filtering weighted matrix w[i,j], and perform symmetric filtering to suppress non-uniform clutter, which is expressed as:

[0085] w[i,j] = e γ·α[i] ·e μ·β[i,j]

[0086] s[i,j] = c[i,j] · w[i,j]

[0087] where γ and μ are the influence factors of the symmetric weight matrices α[i] and β[i,j], and s[i,j] is the signal after symmetric filtering to suppress non-uniform clutter.

[0088] Step 5: Taking the peak signal-to-noise ratio between the echo data before processing and after being processed by Steps 1 - 4 as the objective function, select parameters through the Bayesian optimization algorithm to obtain the parameter values when the maximum peak signal-to-noise ratio is achieved, and finally realize the ringing noise suppression of the ground penetrating radar.

[0089] The Bayesian optimization algorithm is an algorithm for optimizing the objective function. It gradually approaches the optimal value of the objective function by constructing a Gaussian process to select parameters. During the process of constructing the background ringing noise matrix to suppress the ringing noise, some important parameters such as: the sliding window length l, the signal pruning ratio α within the window, the weighting intensity s, and the influence factors γ and μ of the symmetric weight matrix in the symmetric filtering process. The above parameters can be automatically selected through the Bayesian optimization algorithm. Its objective function is the peak signal-to-noise ratio PSNR between the processed data and the pre-processed data in the two processes, and obtain the parameter values when the maximum peak signal-to-noise ratio is achieved through the Bayesian optimization algorithm.

[0090] The present invention proposes a method for suppressing the ringing noise of a ground penetrating radar in a non-uniform environment, which suppresses the ringing noise by constructing a background ringing noise matrix. After reducing the energy of non-uniform clutter using the singular value decomposition method, symmetric filtering is used to suppress non-uniform clutter. For the problem of difficult manual parameter selection, the Bayesian optimization algorithm is used to obtain the parameter values at the maximum peak signal-to-noise ratio with the peak signal-to-noise ratio as the objective function. Finally, the suppression effects of the above method on ringing noise and non-uniform clutter are verified through simulation software.

[0091] Thus, a method for suppressing the ringing noise of a ground penetrating radar in a non-uniform environment is completed. Embodiment

[0092] In order to verify the method for suppressing the ringing noise of a ground penetrating radar in a non-uniform environment proposed by the present invention, a simulation experiment is designed for analysis. The generation of ringing noise and the construction of the non-uniform environment model are as shown in Step 1, where the size of the non-uniform environment model is 1m×0.4m, the grid division size is dx = dy = dz = 0.002m, the upper layer is a Peplinski soil model with a thickness of 0.05m, the lower layer is a dry soil model with a thickness of 0.3m, the proportion of sand in the Peplinski soil model is 0.7, the proportion of clay is 0.3, the range of volumetric water fraction is 0.001 - 0.1, the cross-section of the target is a circle with a radius of 0.02m, the burial depth is 0.15m, the transmitting antenna uses a Ricker wavelet with a frequency of 1.3GHz, the distance between the transmitting and receiving antennas is 0.05m, the step size of antenna translation is 0.005m, and the time window length is 6ns. The schematic diagram of the simulation scenario is as Figure 2 shown.

[0093] Figure 3 (a) is the B-scan image obtained by gprMax simulation. From top to bottom in the original data are the stronger surface reflection echoes, the non-uniform soil model reflection echoes, and the hyperbolic characteristic echoes reflected by the target. Figure 3 (b) is the echo data after adding ringing noise. Figure 3 (c) is the added ringing noise. Affected by the ringing noise, the target echo is completely masked, and it cannot be distinguished from the image whether there is a target underground.

[0094] Figure 4 (a) is the background ringing noise constructed by the method in Step 2 of the present invention. Compared with the Figure 3 (c) added ringing noise, the background ringing noise constructed by this method basically restores the added background ringing noise. Subtracting the obtained background ringing noise from the data in Figure 3 (b) gives the result after ringing noise suppression as shown in Figure 4 (b), where the surface reflection echoes and ringing noise are removed, and the noise suppression effect is good.

[0095] Figure 5 (a) is the result after suppressing the ringing noise and then performing singular value decomposition, where the first two larger singular values are removed, that is Figure 4 the non-uniform clutter horizontal strips with relatively large energy in (b). Figure 5 (b) is Figure 5 based on the singular value decomposition processing of (a) to suppress non-uniform clutter by symmetric filtering. Compared with Figure 4 (b), most of the non-uniform clutter is suppressed and the target echo is obvious.

[0096] The above content is a further detailed description of the present invention in combination with specific implementation manners. It cannot be determined that the specific implementation of the present invention is only limited to these descriptions. For those of ordinary skill in the technical field to which the present invention belongs, without departing from the concept of the present invention, several simple deductions or substitutions can be made, and all should be regarded as belonging to the protection scope of the present invention.

Claims

1. A method for suppressing ringing noise of ground penetrating radar in a non-uniform environment, characterized in that: The specific process is: Step 1: Establish a non-uniform environment model, and simulate ringing noise in the environment model; Step 2: construct a background ringing noise matrix to suppress the ringing noise, and obtain echo data B after removing the ringing noise; Step 3: For the echo data matrix B obtained in step 2, the singular value decomposition method is used to reduce the intensity of the non-uniform clutter to obtain the echo data matrix C; Step 4: For the echo data matrix C, use symmetric filtering to suppress non-uniform clutter and obtain the echo data matrix S; Step 5: Taking the peak signal-to-noise ratio between the echo data before and after processing in steps 1-4 as the objective function, select parameters through the Bayesian optimization algorithm to obtain the parameter value of the echo data matrix at the maximum peak signal-to-noise ratio, and finally achieve ground penetrating radar ringing noise suppression.

2. The method for suppressing ringing noise of ground penetrating radar in a non-uniform environment according to claim 1, characterized in that: The specific method of simulating ringing noise is: in a non-uniform environment model, the receiving antenna and the transmitting antenna are placed at the same horizontal position, and an iron plate is placed at a set distance from the transmitting antenna to reflect the pulse signal emitted by the transmitting antenna. The signal received by the receiving antenna simulates the ringing noise.

3. The method for suppressing ringing noise of ground penetrating radar in a non-uniform environment according to claim 1, characterized in that: The specific process of step 2 is: First, a horizontal window of length l is taken, the echo data matrix is ​​traversed through the sliding window, and the echo data in the window is trimmed and weighted to preliminarily remove the target echo signal to obtain the background ringing noise matrix; Secondly, the generated background ringing noise matrix is ​​iteratively optimized until all target echo signals are removed, leaving only the background ringing noise matrix; Finally, the background ringing noise matrix is ​​subtracted from the original signal echo data to obtain echo data after the ringing noise is removed.

4. The method for suppressing ringing noise of ground penetrating radar in a non-uniform environment according to claim 3, characterized in that: The trimming is as follows: removing the minimum and maximum values ​​in the window according to the trimming ratio α, and averaging the remaining values ​​in the window to obtain the average value A of the window after trimming by the ratio α α ; The weighting is: set the custom parameter s, according to the other values ​​in the window and A α The difference between the two sets the normalized weight, and then the weighted sum is performed to obtain the value A of the background ringing noise in the center of the window rc .

5. The method for suppressing ringing noise of ground penetrating radar in a non-uniform environment according to claim 4, characterized in that: The weighted weight is: Among them, s is a custom parameter, d i is the amplitude a of the i-th element in the window i With A α The difference between Then the weight W i Normalize to get the normalized weight 6. The method for suppressing ringing noise of ground penetrating radar in a non-uniform environment according to claim 1, characterized in that: The method of reducing the intensity of non-uniform clutter by using the singular value decomposition method in step 3 is as follows: for the strip-shaped non-uniform clutter generated by the rough surface, the first one or the first several larger singular values ​​are selected for removal; for the non-uniform clutter generated by underground sand and gravel, the smaller singular value or the singular value close to 0 is selected for removal.

7. The method for suppressing ringing noise of ground penetrating radar in a non-uniform environment according to claim 3, characterized in that: The specific process of step 4 is: sliding the window to traverse all elements in the matrix C, and obtaining the position J with the strongest symmetry in each row of the matrix C m , calculate the strongest position J in each row m Symmetric weight matrix α(i), calculate the symmetric weight matrix β[i,j] at each element in the data matrix C; The influence factors γ and μ of two weights are set, and a synthetic symmetric filter weighting matrix w[i, j] is calculated. The matrix w[i, j] is used to implement symmetric filtering of the matrix C to suppress non-uniform clutter.

8. The method for suppressing ringing noise of ground penetrating radar in a non-uniform environment according to claim 7, characterized in that: The specific process of step 4 is as follows: Calculate each row J m The symmetric weight matrix α(i) at , and normalize it; Where J0(i) represents the column number at the position with the strongest symmetry in the i-th row; Calculate the symmetric weight matrix β[i,j] at each element in the data matrix C and perform normalization; 9. The method for suppressing ringing noise of ground penetrating radar in a non-uniform environment according to claim 7, characterized in that: The step 4 calculates and synthesizes the symmetric filter weighting matrix w[i, j], and uses the matrix w[i, j] to implement symmetric filtering of the matrix C to suppress non-uniform clutter; the specific process is: Calculate the synthetic symmetric filter weight matrix w[i,j] and perform symmetric filtering to suppress non-uniform clutter, expressed as: w[i,j]=e γ·α[i] ·have been μ·β[i,j] s[i,j]=c[i,j]·w[i,j] Among them, γ and μ are the influencing factors of the symmetric weight matrices α[i] and β[i,j], and s[i,j] is the signal after symmetric filtering suppresses non-uniform clutter.

10. The method for suppressing ringing noise of ground penetrating radar in a non-uniform environment according to claim 7, characterized in that: In step 5, the parameters of the sliding window length l, the pruning ratio α, the weighted strength s, and the symmetric weight matrix influence factors γ and μ in the symmetric filtering process are selected through a Bayesian optimization algorithm.

Images