Ground penetrating radar signal denoising method and system based on MFDFA and VMD

Abstract

The invention relates to a ground penetrating radar signal denoising method and system based on MFDFA and VMD. The method comprises the following steps: S1, acquiring ground penetrating radar B-scan original data; s2, using an MFDFA method to calculate parting spectrum width data; s3, determining a reconstruction effective component number according to the width data of the target channel; s4, performing a preset number of VMDs to obtain IMF components; s5, calculating each IMF component by using a DFA method to obtain a Hurst index; s6, when the Hurst index meets the requirement that the number of the target IMF components of the preset index threshold value is larger than or equal to the number of the effective components, the target IMF components are used for reconstruction to obtain a denoised signal of the target channel; and S7, when the number of the effective components is smaller than the number of the effective components, adding 1 on the basis of the preset number, and executing the step S4 and the step S5 until a de-noised signal of the target channel is obtained through reconstruction. And splicing each channel of signals to obtain processed B-scan data, thereby realizing denoising of the ground penetrating radar signals.

Description

Technical Field

[0001] The present invention belongs to the field of exploration geophysical data processing, and particularly relates to a ground penetrating radar signal denoising method and system based on MFDFA and VMD. Background Technique

[0002] Ground Penetrating Radar (GPR) is widely used in many fields such as civil engineering, environmental investigation, archaeology, and mineral resource exploration due to its non-destructive, high-precision, efficient and convenient characteristics. However, it still faces problems such as wide sources and high intensity of signal noise interference.

[0003] In recent years, digital signal processing technology has made important progress in GPR signal noise suppression and signal-to-noise ratio improvement. The wavelet threshold denoising method was earlier applied to non-stationary signals, but the selection of the base wavelet and threshold greatly affects the processing effect of this method. The empirical mode decomposition method has been proposed since 1998 and can adaptively decompose signals into intrinsic mode functions, but there are end effects and mode mixing problems. To solve the above problems, Variational Mode Decomposition (VMD) was proposed, and this method realizes signal decomposition processing by minimizing the variational problem through the alternating direction multiplier method. VMD has been widely used in fields such as biomedicine and mechanical fault detection, and its non-stationary signal decomposition ability superior to empirical mode decomposition demonstrates the application potential of VMD in GPR signal processing.

[0004] For VMD, its effectiveness is affected by the decomposition level K value. If the K value is too small, it will lead to insufficient decomposition, and if it is too large, it will generate false components and waste computational costs. At present, the relevant research on the selection of K is limited. Dragomiretskiy and Zosso published a paper stating that the K value is adjusted through modal spectrum overlap analysis, and Lahmiri et al. published a paper setting the decomposition level K value of VMD to be the same as that of EMD in image denoising. However, the above methods rely on the experience of operators and are not applicable to large-scale multi-channel two-dimensional signals including GPR signals.

[0005] How to efficiently and adaptively denoise GPR signals has become a technical problem that urgently needs to be solved. Summary of the Invention

[0006] The purpose of the present invention is to provide a ground penetrating radar signal denoising method and system based on MFDFA and VMD for the problem that ground penetrating radar signals are overwhelmed by strong background noise and clutter interference and the signal-to-noise ratio of ground penetrating radar B-scan data is low, so as to suppress noise interference in low-signal-to-noise-ratio ground penetrating radar B-scan data and improve the signal-to-noise ratio of the data.

[0007] To solve the above technical problems, the technical solution adopted by the present invention is as follows:

[0008] In a first aspect, a ground penetrating radar signal denoising method based on MFDFA and VMD is provided, including:

[0009] S1, obtaining the original B-scan data of the ground penetrating radar, where the original B-scan data includes multiple channels of original signals;

[0010] S2, using the multifractal detrended fluctuation analysis (MFDFA) method to calculate the width data of the multifractal spectrum of each channel of the original signal;

[0011] S3, determining the number of effective components for reconstructing the original signal of the target channel according to the width data of the target channel;

[0012] S4, performing variational mode decomposition (VMD) on the original signal of the target channel for a preset number of times to obtain a preset number of intrinsic mode function (IMF) components;

[0013] S5, using the detrended fluctuation analysis (DFA) method to calculate the Hurst exponent of each IMF component respectively;

[0014] S6, when the number of target IMF components whose Hurst exponent satisfies the preset exponent threshold is greater than or equal to the number of effective components, using the target IMF components for reconstruction to obtain the denoised signal of the target channel;

[0015] S7, when the number of target IMF components whose Hurst exponent satisfies the preset exponent threshold is less than the number of effective components, adding 1 to the preset number, and executing steps S4 and S5 until the denoised signal of the target channel is reconstructed.

[0016] Further, the original signal is m channels, and m is a positive integer greater than or equal to 1.

[0017] Further, step S3, determining the number of effective components for reconstructing the original signal of the target channel according to the width data of the target channel, includes:

[0018] Comparing the width data Δα of the multifractal spectrum of the original signal of the target channel j with a first preset threshold E1 and a second preset threshold E2, where E1 is less than E2;

[0019] When Δα < E1, determining that the value of the number of effective components J of the original signal of the target channel j is J1;

[0020] When E1 < Δα < E2, determining that the value of the number of effective components J of the original signal of the target channel j is J2;

[0021] When Δα > E2, determining that the value of the number of effective components J of the original signal of the target channel j is J3.

[0022] Further, in step S4, performing variational mode decomposition (VMD) on the original signal of the target trace for a preset number of times to obtain a preset number of intrinsic mode function (IMF) components, including:

[0023] Obtaining the preset number K according to the number of effective components J plus 1, where K = J + 1;

[0024] Performing VMD on the original signal of the target trace j for the preset number K times to obtain K IMF components.

[0025] Further, after step S5, it further includes:

[0026] Comparing the Hurst exponent of each IMF component with a preset exponent threshold respectively;

[0027] Taking the IMF components whose corresponding Hurst exponent is greater than or equal to the preset exponent threshold as target IMF components;

[0028] Taking the IMF components whose corresponding Hurst exponent is less than the preset exponent threshold as non-target IMF components.

[0029] Further, after taking the IMF components whose corresponding Hurst exponent is greater than or equal to the preset exponent threshold as target IMF components, it further includes:

[0030] Counting the number of target IMF components and comparing the number of target IMF components with the number of effective components J.

[0031] Second, a ground penetrating radar signal denoising system based on MFDFA and VMD is provided, including:

[0032] A data acquisition module, a multifractal detrended fluctuation analysis (MFDFA) module, an effective component number calculation module, a variational mode decomposition (VMD) module, a Hurst exponent calculation module, and a signal reconstruction module;

[0033] The data acquisition module is used to acquire the original ground penetrating radar B-scan data, and the original B-scan data includes multi-trace original signals;

[0034] The MFDFA module is used to calculate the width data of the multifractal spectrum of each trace's original signal by using the MFDFA method;

[0035] The effective component number calculation module is used to determine the number of effective components for reconstructing the original signal of the corresponding target trace according to the width data of the target trace;

[0036] The VMD module is used to perform VMD on the original signal of the target trace for a preset number of times to obtain a preset number of intrinsic mode function (IMF) components;

[0037] The Hurst exponent calculation module is used to calculate the Hurst exponent of each IMF component respectively by using the detrended fluctuation analysis (DFA) method for each IMF component;

[0038] The signal reconstruction module is used to reconstruct the denoised signal of the target trace by using the target IMF components when the number of target IMF components whose Hurst exponents satisfy the preset exponent threshold is greater than or equal to the number of effective components;

[0039] The VMD module is also used to, when the number of target IMF components whose Hurst exponents satisfy the preset exponent threshold is less than the number of effective components, after adding 1 on the basis of the preset number, then perform VMD to obtain IMF components with the preset number plus 1, and use the Hurst exponent calculation module to calculate the Hurst exponents of each IMF component until the denoised signal of the target trace is reconstructed.

[0040] Further, the original signal is m traces, where m is a positive integer greater than or equal to 1.

[0041] The effective component number calculation module is specifically used to compare the width data Δα of the multifractal spectrum of the original signal of the target trace j with a first preset threshold E1 and a second preset threshold E2, where E1 < E2; when Δα < E1, determine that the value of the number of effective components J of the original signal of the target trace j is J1; when E1 < Δα < E2, determine that the value of the number of effective components J of the original signal of the target trace j is J2; when Δα > E2, determine that the value of the number of effective components J of the original signal of the target trace j is J3.

[0042] Further, the VMD module is specifically used to obtain a preset number K according to the number of effective components J plus 1, K = J + 1; perform VMD with the preset number K on the original signal of the target trace j to obtain K IMF components.

[0043] Further, the ground penetrating radar signal denoising system further includes a target IMF component extraction module.

[0044] The target IMF component extraction module is used to compare the Hurst exponents of each IMF component with the preset exponent threshold respectively; take the IMF components whose corresponding Hurst exponents are greater than or equal to the preset exponent threshold as the target IMF components; take the IMF components whose corresponding Hurst exponents are less than the preset exponent threshold as the non-target IMF components.

[0045] The beneficial effects achieved by the present invention:

[0046] S1. Obtain the original GPR B-scan data, where the original B-scan data includes multiple channels of original signals; S2. Use the multifractal detrended fluctuation analysis (MFDFA) method to calculate the width data of the multifractal spectrum for each channel of the original signal; S3. Determine the number of effective components for reconstructing the original signal of the target channel according to the width data of the target channel; S4. Perform a preset number of variational mode decompositions (VMD) on the original signal of the target channel to obtain a preset number of intrinsic mode function (IMF) components; S5. Use the detrended fluctuation analysis (DFA) method to calculate the Hurst exponent for each IMF component respectively; S6. When the number of target IMF components whose Hurst exponent meets the preset exponent threshold is greater than or equal to the number of effective components, use the target IMF components for reconstruction to obtain the denoised signal of the target channel; S7. When the number of target IMF components whose Hurst exponent meets the preset exponent threshold is less than the number of effective components, increment by 1 on the basis of the preset number, and execute steps S4 and S5 until the denoised signal of the target channel is reconstructed.

[0047] By introducing MFDFA, the adaptive selection of the K value and component screening and reconstruction of VMD decomposition are realized, so as to achieve the denoising of GPR signals, and the effects of suppressing background noise and improving the signal-to-noise ratio of the original signal are obtained. Brief Description of the Drawings

[0048] Figure 1 It is a flowchart of the GPR signal denoising method based on MFDFA and VMD of the present invention;

[0049] Figure 2 It is a lining model diagram of the present invention;

[0050] Figure 3 It is a simulated simulation data diagram of the lining model of the present invention;

[0051] Figure 4 It is a radar image with synthetic noise of the present invention;

[0052] Figure 5 It is a graph of the width values of the signal classification spectra of each channel of the present invention;

[0053] Figure 6 It is a Hurst exponent graph of the present invention with K values from 17 to 23;

[0054] Figure 7 It is a result radar image of the present invention for processing synthetic noise;

[0055] Figure 8 It is a structural diagram of the GPR signal denoising system based on MFDFA and VMD of the present invention. Detailed Embodiments

[0056] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention and cannot be used to limit the protection scope of the present invention.

[0057] As Figure 1 shown, an embodiment of the present invention provides a ground penetrating radar signal denoising method based on MFDFA and VMD, including:

[0058] S1, obtaining the original B-scan data of the ground penetrating radar, where the original B-scan data includes multiple channels of original signals;

[0059] In this embodiment, taking Figure 2 a lining model with a length of 4 m and a depth of 1.1 m shown as an example, where the uppermost layer is an air layer with a thickness of 0.1, the second layer is a secondary lining with a thickness of 0.4 m, the third layer is a primary support with a thickness of 0.2 m, and the lowermost layer is a surrounding rock with a thickness of 0.4 m. There are two square cavities and two square water-containing anomalies in the surrounding rock and the primary support. With the center frequency of the ground penetrating radar antenna being 900 MHz, simulation is carried out to obtain the Figure 3 simulation data diagram of the subgrade model shown, the scan channel numbers are from 1 to 380, a total of 380 channels, the time window is 20 ns, and the number of sampling points is 3393 points; Synthetic noise with Gaussian noise as the background, including specific frequency crosstalk signals and sharp pulse signal interference, is added to the obtained simulation data, and the signal-to-noise ratio is about -15 dB, forming the Figure 4 radar image with synthetic noise shown. It can be seen from the Figure 4 radar image that the noise interference is serious. The reflections of the upper and lower interfaces of the water-containing anomaly in the 140th channel are almost submerged by the noise, and the reflection of the lower interface of the cavity anomaly in the 240th channel is also almost invisible. At the same time, the boundary of the layered medium is also blurred and difficult to distinguish.

[0060] Taking Figure 4 the radar image in it as an example, the original signals in the original B-scan data are m = 380 channels.

[0061] S2, using the MFDFA method to calculate the width data of the multifractal spectrum of each channel of the original signal;

[0062] In this embodiment, the MFDFA method is used for calculation. The width data Δα of the multifractal spectrum is defined as the difference between the maximum value and the minimum value of α (Δα = α_max - α_min), which characterizes the diversity of signal singularity. The obtained Figure 5 Δα data diagram of each channel signal shown, the high Δα positions are respectively near the 50th channel, 140th channel, 240th channel, and 325th channel, corresponding to the positions of the abnormal bodies.

[0063] S3. Determine the number of effective components for reconstructing the original signal corresponding to the target trace according to the width data of the target trace;

[0064] In this embodiment, the number of effective components J required for reconstructing the signal is determined according to the width data Δα of the multifractal spectrum. The key lies in that Δα reflects the distribution range of the singularity exponents in the signal; the larger Δα is, the more complex the singular structure contained in the signal is, and more components are needed to cover these different singularities. The number of effective components J is in a proportional relationship with Δα; the first preset threshold E1 and the second preset threshold E2 are set according to the signal characteristics (such as the wavelet scale limit) or the decomposition method, and E1 is less than E2;

[0065] Compare the width data Δα of the multifractal spectrum of the original signal of the target trace j with the first preset threshold E1 and the second preset threshold E2, where E1 is less than E2;

[0066] When Δα < E1, at this time, the distribution range of the signal singularity is relatively narrow and the complexity is relatively low. Determine that the value of the number of effective components J of the original signal of the target trace j is J1;

[0067] When E1 < Δα < E2, determine that the value of the number of effective components J of the original signal of the target trace j is J2;

[0068] When Δα > E2, determine that the value of the number of effective components J of the original signal of the target trace j is J3;

[0069] Generally, the values of J1, J2, and J3 are 1, 2, 3 or 2, 3, 4, and can be selected according to the complexity of the ground penetrating radar target.

[0070] S4. Perform a preset number of VMDs on the original signal of the target trace to obtain a preset number of IMF components;

[0071] In this embodiment, taking the 50th trace of the B-scan original data as an example, according to the above step S3, determine that the number of effective components J required for reconstructing this trace is J = 3;

[0072] Obtain the preset number K according to the number of effective components J plus 1, K = J + 1 = 4;

[0073] Perform VMD with the preset number K value of 4 on the original signal of the 50th target trace to obtain 4 IMF components.

[0074] S5. Use the DFA method to calculate the Hurst exponent of each IMF component respectively;

[0075] In this embodiment, calculate each IMF component according to the number of IMF components in step S4 to obtain the Hurst exponent of each IMF component; for example Figure 6As shown, it is a Hurst index graph with K values ranging from 17 to 23 and a preset exponential threshold of 0.8;

[0076] Compare the Hurst indices of each IMF component with the preset exponential threshold respectively;

[0077] Take the IMF components whose corresponding Hurst indices are greater than or equal to the preset exponential threshold as the target IMF components;

[0078] Take the IMF components whose corresponding Hurst indices are less than the preset exponential threshold as non-target IMF components;

[0079] Count the number of target IMF components and compare the number of target IMF components with the number of effective components J;

[0080] According to Figure 6 As shown, when the K value is from 17 to 19, only two IMF components satisfy the Hurst index greater than 0.8, that is, there is only one target IMF component, which does not reach the required number of effective components 3 for reconstruction;

[0081] The condition is satisfied when the K value is 20, and compared with the K values from 21 to 23, the Hurst scales of each component are not very different, which means that each component has been effectively separated when the K value is from 20 to 23. Therefore, choosing the K value of 20 can not only ensure the full decomposition of the signal of this trace, but also improve the decomposition efficiency and save the calculation cost, while avoiding the generation of false components due to over-decomposition.

[0082] S6. When the number of target IMF components whose Hurst index satisfies the preset exponential threshold is greater than or equal to the number of effective components, use the target IMF components for reconstruction to obtain the denoised signal of the target trace;

[0083] In this embodiment, when the K value is 20, the number of target IMF components is greater than or equal to the number of effective components, and the target IMF components are used for reconstruction to obtain the denoised signal of the target trace.

[0084] S7. When the number of target IMF components whose Hurst index satisfies the preset exponential threshold is less than the number of effective components, add 1 on the basis of the preset number and return to execute step S4.

[0085] In this embodiment, assume that when the K value is 4, the number of target IMF components is 1, which is less than the number of effective components 2. At this time, it is necessary to make K + 1 = 4, and then execute step S4 and step S5 until K = 20.

[0086] After completing the above steps for all trace signals, the results are as Figure 7 shown. The results show that the reflections of the upper and lower interfaces of all abnormal bodies are clearly distinguishable after processing, and the interfaces of the layered media are also significantly enhanced.

[0087] Beneficial effects achieved by the present invention:

[0088] S1. Obtain the original GPR B-scan data, where the original B-scan data includes multiple channels of original signals; S2. Use the multifractal detrended fluctuation analysis (MFDFA) method to calculate the width data of the multifractal spectrum of each channel of the original signal; S3. Determine the number of effective components for reconstructing the original signal of the target channel according to the width data of the target channel; S4. Perform variational mode decomposition (VMD) on the original signal of the target channel for a preset number of times to obtain a preset number of intrinsic mode function (IMF) components; S5. Use the detrended fluctuation analysis (DFA) method to calculate the Hurst exponent of each IMF component respectively; S6. When the number of target IMF components whose Hurst exponent satisfies the preset exponent threshold is greater than or equal to the number of effective components, use the target IMF components for reconstruction to obtain the denoised signal of the target channel; S7. When the number of target IMF components whose Hurst exponent satisfies the preset exponent threshold is less than the number of effective components, increment by 1 on the basis of the preset number, and execute steps S4 and S5 until the denoised signal of the target channel is reconstructed.

[0089] By introducing MFDFA, the adaptive selection of the K value for VMD decomposition and component screening and reconstruction are realized, so as to denoise the GPR signal, achieving the effects of suppressing background noise and improving the signal-to-noise ratio of the original signal.

[0090] Combined with the GPR signal denoising method based on MFDFA and VMD described in the above embodiments, the GPR signal denoising system based on MFDFA and VMD will be described below through embodiments.

[0091] As Figure 8 shown, an embodiment of the present invention provides a GPR signal denoising system based on MFDFA and VMD, including:

[0092] A data acquisition module 801, a multifractal detrended fluctuation analysis (MFDFA) module 802, an effective component number calculation module 803, a variational mode decomposition (VMD) module 804, a Hurst exponent calculation module 805, and a signal reconstruction module 806;

[0093] The data acquisition module 801 is used to obtain the original GPR B-scan data, where the original B-scan data includes multiple channels of original signals;

[0094] The MFDFA module 802 is used to calculate the width data of the multifractal spectrum of each channel of the original signal by using the MFDFA method;

[0095] The effective component number calculation module 803 is used to determine the number of effective components for reconstructing the original signal of the target channel according to the width data of the target channel;

[0096] The VMD module 804 is configured to perform a preset number of VMDs on the original signal of the target channel to obtain a preset number of Intrinsic Mode Function (IMF) components;

[0097] The Hurst exponent calculation module 805 is configured to calculate the Hurst exponent of each IMF component respectively by using the Detrended Fluctuation Analysis (DFA) method for each IMF component;

[0098] The signal reconstruction module 806 is configured to, when the number of target IMF components whose Hurst exponents meet the preset exponent threshold is greater than or equal to the number of effective components, use the target IMF components for reconstruction to obtain the denoised signal of the target channel;

[0099] The VMD module 804 is further configured to, when the number of target IMF components whose Hurst exponents meet the preset exponent threshold is less than the number of effective components, after adding 1 to the preset number, perform VMD again to obtain IMF components with the preset number plus 1, and use the Hurst exponent calculation module 805 to calculate the Hurst exponent of each IMF component until the denoised signal of the target channel is reconstructed.

[0100] Preferably, the original signal is m channels, where m is a positive integer greater than or equal to 1.

[0101] The effective component number calculation module 803 is specifically configured to compare the width data Δα of the multifractal spectrum of the original signal of the target channel j with a first preset threshold E1 and a second preset threshold E2, where E1 < E2; when Δα < E1, determine that the value of the effective component number J of the original signal of the target channel j is J1; when E1 < Δα < E2, determine that the value of the effective component number J of the original signal of the target channel j is J2; when Δα > E2, determine that the value of the effective component number J of the original signal of the target channel j is J3.

[0102] Preferably, the VMD module 804 is specifically configured to obtain a preset number K according to the effective component number J plus 1, where K = J + 1; perform VMD on the original signal of the target channel j for the preset number K to obtain K IMF components.

[0103] Preferably, the ground penetrating radar signal denoising system further includes a target IMF component extraction module.

[0104] The target IMF component extraction module is configured to compare the Hurst exponents of each IMF component with a preset exponent threshold respectively; use the IMF components whose corresponding Hurst exponents are greater than or equal to the preset exponent threshold as target IMF components; use the IMF components whose corresponding Hurst exponents are less than the preset exponent threshold as non-target IMF components.

[0105] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0106] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, as well as the combination of flows and / or blocks in the flowchart and / or block diagram, can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0107] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0108] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0109] The above are only embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention are included in the scope of the claims of the present invention pending approval.

Claims

1. A ground penetrating radar signal denoising method based on MFDFA and VMD, characterized in that Including: S1. Obtain the original GPR B-scan data, where the B-scan original data includes multiple channels of original signals; S2. Use the Multifractal Detrended Fluctuation Analysis (MFDFA) method to calculate the width data of the multifractal spectrum of each channel of the original signal; S3. Determine the number of effective components for reconstructing the original signal of the corresponding target channel according to the width data of the target channel; S4. Perform a preset number of Variational Mode Decomposition (VMD) on the original signal of the target channel to obtain the preset number of Intrinsic Mode Function (IMF) components; S5. Use the Detrended Fluctuation Analysis (DFA) method to calculate the Hurst exponent of each IMF component respectively; S6. When the number of target IMF components whose Hurst exponent meets the preset exponent threshold is greater than or equal to the number of effective components, use the target IMF components for reconstruction to obtain the denoised signal of the target channel; S7. When the number of target IMF components whose Hurst exponent meets the preset exponent threshold is less than the number of effective components, increment by 1 based on the preset number, and execute steps S4 and S5 until the denoised signal of the target channel is reconstructed.

2. The ground penetrating radar signal denoising method based on MFDFA and VMD according to claim 1, characterized in that The original signal is m channels, where m is a positive integer greater than or equal to 1.

3. The ground penetrating radar signal denoising method based on MFDFA and VMD according to claim 2, wherein The step S3, determining the number of effective components for reconstructing the original signal of the corresponding target channel according to the width data of the target channel, includes: Compare the width data Δα of the multifractal spectrum of the original signal of target channel j with a first preset threshold E1 and a second preset threshold E2, where E1 is less than E2; When Δα < E1, determine that the value of the number of effective components J of the original signal of target channel j is J1; When E1 < Δα < E2, determine that the value of the number of effective components J of the original signal of target channel j is J2; When Δα > E2, determine that the value of the number of effective components J of the original signal of target channel j is J3.

4. The ground penetrating radar signal denoising method based on MFDFA and VMD according to claim 3, wherein The step S4, performing a preset number of VMD on the original signal of the target channel to obtain the preset number of IMF components, includes: Obtain a preset number K according to the number of effective components J plus 1, where K = J + 1; Perform the preset number K of VMD on the original signal of target channel j to obtain K IMF components.

5. The ground penetrating radar signal denoising method based on MFDFA and VMD according to claim 4, characterized in that, After the step S5, it further includes: Compare the Hurst exponents of each IMF component with a preset exponent threshold respectively; Take the IMF components whose corresponding Hurst exponents are greater than or equal to the preset exponent threshold as target IMF components; Take the IMF components whose corresponding Hurst exponents are less than the preset exponent threshold as non-target IMF components.

6. The ground penetrating radar signal denoising method based on MFDFA and VMD according to claim 5, characterized in that, After taking the IMF components whose corresponding Hurst exponents are greater than or equal to the preset exponent threshold as target IMF components, it further includes: Count the number of target IMF components, and compare the number of target IMF components with the number of effective components J.

7. A ground penetrating radar signal denoising system based on MFDFA and VMD, characterized in that, Including: A data acquisition module, a multifractal detrended fluctuation analysis (MFDFA) module, an effective component number calculation module, a variational mode decomposition (VMD) module, a Hurst exponent calculation module, and a signal reconstruction module; The data acquisition module is used to acquire the original ground penetrating radar B-scan data, and the original B-scan data includes multiple channels of original signals; The MFDFA module is used to calculate the width data of the multifractal spectrum of each channel of the original signal using the MFDFA method; The effective component number calculation module is used to determine the number of effective components for reconstructing the original signal of the target channel according to the width data of the target channel; The VMD module is used to perform VMD on the original signal of the target channel for a preset number of times to obtain the preset number of intrinsic mode function (IMF) components; The Hurst exponent calculation module is used to calculate the Hurst exponent of each IMF component respectively by using the detrended fluctuation analysis (DFA) method for each IMF component; The signal reconstruction module is used to reconstruct the denoised signal of the target channel by using the target IMF components when the number of target IMF components whose Hurst exponent meets the preset exponent threshold is greater than or equal to the number of effective components; The VMD module is further used to, when the number of target IMF components whose Hurst exponent meets the preset exponent threshold is less than the number of effective components, after adding 1 to the preset number, perform VMD again to obtain the IMF components with the preset number plus 1, and use the Hurst exponent calculation module to calculate the Hurst exponent of each IMF component until the denoised signal of the target channel is reconstructed.

8. The ground penetrating radar signal denoising system based on MFDFA and VMD according to claim 7, characterized in that, The original signals are m channels, where m is a positive integer greater than or equal to 1, The effective component number calculation module is specifically used to compare the width data Δα of the multifractal spectrum of the original signal of the target channel j with a first preset threshold E1 and a second preset threshold E2, where E1 is less than E2; when Δα < E1, it is determined that the value of the number of effective components J of the original signal of the target channel j is J1; when E1 < Δα < E2, it is determined that the value of the number of effective components J of the original signal of the target channel j is J2; when Δα > E2, it is determined that the value of the number of effective components J of the original signal of the target channel j is J3.

9. The ground penetrating radar signal denoising system based on MFDFA and VMD according to claim 8, wherein, The VMD module is specifically used to obtain a preset number K according to the number of effective components J plus 1, where K = J + 1; perform VMD on the original signal of the target channel j for the preset number K times to obtain K IMF components.

10. The ground penetrating radar signal denoising system based on MFDFA and VMD according to claim 9, characterized in that, The ground penetrating radar signal denoising system further includes a target IMF component extraction module, The target IMF component extraction module is used to compare the Hurst exponents of the respective IMF components with a preset exponent threshold; the IMF components with corresponding Hurst exponents greater than or equal to the preset exponent threshold are used as target IMF components; and the IMF components with corresponding Hurst exponents less than the preset exponent threshold are used as non-target IMF components.

Images