Grounding grid sound signal filtering and noise reduction method and device based on Fourier transform disassembly and parallel acceleration

Abstract

The invention relates to the technical field of power distribution networks, and discloses a grounding grid sound signal filtering and noise reduction method and device based on Fourier transform disassembly and parallel acceleration, and the method comprises the steps: obtaining a mixed multi-sound-source signal; segmenting the multi-sound-source signal into grounding grid echo signal segmentation sub-frequencies with different center frequencies; performing Fourier transform disassembly and parallel acceleration on the grounding grid echo signal segmentation sub-frequency to generate a grounding grid echo signal segmentation sub-frequency parallel composition factor; and carrying out noise reduction processing on the sub-frequency parallel factors, and combining to obtain a global noise reduction signal. According to the method, the multi-sound-source signal is divided into the grounding grid echo signal segmentation sub-frequencies with different center frequencies, the filtering noise reduction rate is improved through Fourier transform disassembly and parallel acceleration, the echo signal quality of the grounding grid corrosion state in the actual environment is improved, and a reliable and stable data basis is provided for observation of the grounding grid corrosion state.

Description

Technical Field

[0001] The present invention relates to the technical field of power distribution networks, and in particular to a method and device for filtering and denoising grounding network acoustic signals based on Fourier transform disassembly and parallel acceleration. Background Art

[0002] The grounding grid status acoustic signal refers to the echo signal generated by transmitting square wave pulses to the grounding grid during grounding grid status monitoring. Because the echo signal carries information about the grounding grid's status, analyzing and processing these signals can help power grid personnel understand the corrosion status of the grounding grid.

[0003] However, during on-site testing, various electromagnetic interferences exist around the grounding grid, such as stray currents from DC equipment in high-speed rail stations, overvoltages from radio waves and lightning, and other sources. These interferences can reduce the signal-to-noise ratio of the echo signal and affect the accuracy of acoustic signal classification. Therefore, improvements to existing technologies are needed.

[0004] The above information is presented as background information only to assist with an understanding of the present disclosure and is not a determination or admission that any of the above may be applicable as prior art with respect to the present disclosure. Summary of the Invention

[0005] The present invention provides a grounding grid acoustic signal filtering and noise reduction method and device based on Fourier transform disassembly and parallel acceleration to solve the problems existing in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solutions:

[0007] A grounding grid acoustic signal filtering and noise reduction method based on Fourier transform disassembly and parallel acceleration, comprising:

[0008] Obtain mixed multi-source signals;

[0009] Splitting the multi-source signal into ground grid echo signal split sub-frequency components with different center frequencies;

[0010] Performing Fourier transform decomposition and parallel acceleration on the ground grid echo signal divided sub-frequency to generate parallel component factors of the ground grid echo signal divided sub-frequency;

[0011] Noise reduction processing is performed on the sub-frequency parallel factors, and a global noise reduction signal is obtained after merging.

[0012] Optionally, dividing the multiple sound source signals into ground grid echo signal segmentation sub-frequency components with different center frequencies includes:

[0013] The low-pass filter is moved to different center frequencies through complex modulation method to generate multiple band-pass filter banks;

[0014] The input multi-source signal is subjected to multi-phase decomposition, modulated to a baseband, and divided into a plurality of ground grid echo signal sub-frequency divisions with limited bandwidth and different center frequencies.

[0015] Optionally, the performing Fourier transform decomposition and parallel acceleration on the sub-frequency division of the ground grid echo signal includes:

[0016] Get the ground grid echo signal split sub-frequency , calculate the ground grid echo signal segmentation sub-frequency in Fourier domain Fourier transform of ;in express The continuous Fourier transform of and is the continuous frequency;

[0017] right Sampling is performed to obtain a two-dimensional array , realizing the conversion from continuous domain to discrete domain; the two-dimensional array By having A finite square of data points Sampling is performed, in which and , and is a discrete frequency, for , ;

[0018] right Calculate the forward discrete Fourier transform to obtain its frequency domain representation;

[0019] Using the two-dimensional Fourier transform to decompose the operator, Multiply , and converted back to the discrete domain via backward discrete Fourier transform;

[0020] Will Disassembled into , transform the two-dimensional problem into two one-dimensional problems, and get and Two parallel component factors.

[0021] Optionally, performing noise reduction processing on the sub-frequency parallel factors includes:

[0022] Each grounding grid echo signal is divided into sub-frequency parallel component factors and multi-phase low-pass filtering is performed to filter out high-frequency interference noise;

[0023] The filtered ground grid echo signal is divided into sub-frequency components and modulated back to the original center frequency in parallel.

[0024] Optionally, the ground grid acoustic signal filtering and noise reduction method based on Fourier transform decomposition and parallel acceleration further includes:

[0025] Based on the following formula, the global noise reduction signal is iteratively optimized through the constrained variation decomposition function:

[0026] ;

[0027] in, is the Fourier transform decomposition operator (second-order difference), λ and μ are weight parameters that control the smoothing and sparsity strength;

[0028] It is the fidelity item. is a smoothness constraint, Sparse constraints.

[0029] Optionally, the iterative optimization includes:

[0030] The accelerated gradient descent method is used to solve the denoised signal ydenoised that satisfies the constraints based on the following gradient descent formula:

[0031] ;

[0032] ;

[0033] Where, is the fourth-order difference (the second-order derivative of the smoothing term), and η is the learning rate.

[0034] The present invention also provides a grounding grid acoustic signal filtering and noise reduction device based on Fourier transform disassembly and parallel acceleration, comprising:

[0035] a multi-band decomposition module configured to split the multi-source signal into ground grid echo signal segmentation sub-frequency components with different center frequencies;

[0036] A Fourier transform decomposition parallel acceleration module is configured to perform two-dimensional Fourier transform decomposition and GPU parallel acceleration calculation on the sub-frequency segmentation of the ground grid echo signal to generate parallel component factors of the sub-frequency segmentation of the ground grid echo signal;

[0037] The noise reduction processing module is configured to perform multiphase low-pass filtering and signal merging on the parallel component factors of the ground grid echo signal segmentation sub-frequency to obtain a global noise reduction signal.

[0038] Optionally, the grounding grid acoustic signal filtering and noise reduction device based on Fourier transform disassembly and parallel acceleration further includes:

[0039] The constrained variation decomposition module is configured to iteratively optimize the global denoised signal through an energy function.

[0040] Optionally, the multi-band decomposition module includes:

[0041] RLC circuit is used to correct the excessive mutation of multi-source signals within the constraint range;

[0042] The complex modulation unit is used to move the low-pass filter to different center frequencies to generate a band-pass filter bank.

[0043] Optionally, the Fourier transform disassembly parallel acceleration module includes:

[0044] GPU computing unit for performing two one-dimensional Fourier transform and complex multiplication operations in parallel;

[0045] The data transposition unit is used to separate the frequency domain components in the horizontal and vertical directions.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] The present invention provides a method and device for filtering and denoising grounding grid acoustic signals based on Fourier transform decomposition and parallel acceleration. The method divides the multi-source signal into grounding grid echo signal segmentation sub-frequency segments with different center frequencies, and performs Fourier transform decomposition and parallel acceleration to improve the filtering and denoising rate, thereby improving the echo signal quality of the grounding grid corrosion state in actual environments and providing a reliable and stable data basis for observing the grounding grid corrosion state.

[0048] The present invention has other features and advantages that will be apparent from or will be described in detail in the accompanying drawings and the following detailed description incorporated herein, which together serve to explain certain principles of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0050] Figure 1 This is a flow chart of a method for filtering and denoising grounding grid acoustic signals based on Fourier transform disassembly and parallel acceleration, provided in the first embodiment of the present invention;

[0051] Figure 2 This is a structural diagram of an RLC circuit involved in a grounding grid acoustic signal filtering and noise reduction method based on Fourier transform disassembly and parallel acceleration provided in Example 1 of the present invention;

[0052] Figure 3 This is a flowchart of step S2 in a method for filtering and denoising ground grid acoustic signals based on Fourier transform disassembly and parallel acceleration provided in Example 1 of the present invention;

[0053] Figure 4 This is a flowchart of step S5 in a method for filtering and denoising ground grid acoustic signals based on Fourier transform disassembly and parallel acceleration provided in Example 1 of the present invention;

[0054] Figure 5 This is a structural diagram of a grounding grid acoustic signal filtering and noise reduction device based on Fourier transform disassembly and parallel acceleration provided by the second embodiment of the present invention;

[0055] Figure 6 This is a structural diagram of a multi-band decomposition module in a grounding grid acoustic signal filtering and noise reduction device based on Fourier transform disassembly and parallel acceleration provided by the second embodiment of the present invention;

[0056] Figure 7 It is a structural diagram of a Fourier transform disassembly and parallel acceleration module in a grounding grid acoustic signal filtering and noise reduction device based on Fourier transform disassembly and parallel acceleration provided in the second embodiment of the present invention.

[0057] Figure numerals: 10, multi-band decomposition module; 101, RLC circuit; 102, complex modulation unit; 20, Fourier transform decomposition parallel acceleration module; 201, GPU computing unit; 202, data transposition unit; 30, noise reduction processing module; 40, constrained variation decomposition module. DETAILED DESCRIPTION

[0058] In order to explain in detail the possible application scenarios, technical principles, specific solutions that can be implemented, and the purpose and effects of this application, the following is a detailed description of the specific embodiments listed in conjunction with the accompanying drawings. The embodiments described herein are only used to more clearly illustrate the technical solutions of this application and are therefore only examples and are not intended to limit the scope of protection of this application.

[0059] References to "embodiments" herein mean that the specific features, structures, or characteristics described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the word "embodiment" in various places in the specification does not necessarily refer to the same embodiment, nor does it particularly limit its independence or relevance to other embodiments. In principle, in this application, as long as there are no technical contradictions or conflicts, the various technical features mentioned in the embodiments can be combined in any manner to form a corresponding implementable technical solution.

[0060] Unless otherwise defined, the technical terms used herein have the same meanings as those generally understood by those skilled in the art to which this application belongs; the use of relevant terms herein is only for describing specific embodiments and is not intended to limit this application.

[0061] In the description of this application, the term "and / or" is used to describe a logical relationship between objects, indicating that three relationships can exist. For example, A and / or B means: A exists, B exists, and both A and B exist. In addition, the character " / " in this document generally indicates that the objects before and after are in a logical "or" relationship.

[0062] In this application, terms such as "first" and "second" are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual quantity, priority or sequence relationship between these entities or operations.

[0063] Without further limitations, in this application, the words "include", "comprise", "have" or other similar expressions used in the sentences are intended to cover non-exclusive inclusion. These expressions do not exclude the presence of additional elements in the process, method or product including the elements, so that the process, method or product including a series of elements may include not only those defined elements, but also other elements not explicitly listed, or elements inherent to such process, method or product.

[0064] Consistent with the understanding in the Examination Guidelines, in this application, expressions such as "greater than," "less than," and "exceed" are understood to exclude the number itself; expressions such as "above," "below," and "within" are understood to include the number itself. Furthermore, in the description of the embodiments of this application, "multiple" means more than two (including two), and similar expressions related to "multiple" are also understood in this manner, such as "multiple groups," "multiple times," etc., unless otherwise specifically defined.

[0065] In the description of the embodiments of the present application, the space-related expressions used, such as "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "vertical", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc., indicate the orientation or position relationship based on the orientation or position relationship shown in the specific embodiments or drawings, and are only for the convenience of describing the specific embodiments of the present application or facilitating the reader's understanding, and do not indicate or imply that the device or component referred to must have a specific position, a specific orientation, or be constructed or operated in a specific orientation. Therefore, it should not be understood as a limitation on the embodiments of the present application.

[0066] Unless otherwise expressly specified or limited, in the description of the embodiments of the present application, the terms "installed", "connected", "connected", "fixed", "set", etc. used should be understood in a broad sense. For example, the "connection" can be a fixed connection, a detachable connection, or an integrated setting; it can be a mechanical connection, an electrical connection, or a communication connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be the internal connection of two elements or the interaction relationship between two elements. For those skilled in the art of the present application, the specific meanings of the above terms in the embodiments of the present application can be understood according to the specific circumstances.

[0067] In order to solve the problem of interference of the grounding grid environment on the echo signal, and considering the premise of a huge amount of sample data to be processed, the present invention proposes a grounding grid acoustic signal filtering and denoising device based on constrained variation decomposition and Fourier transform disassembly and parallel acceleration. Based on the micro-hardware denoising device, the range of the multi-source echo signal with excessive mutation is corrected, the grounding grid acoustic signal is segmented by a multi-band decomposition module, and the denoising operation is completed through the constrained variation decomposition algorithm. A Fourier transform disassembly parallel acceleration architecture is proposed to improve the model filtering and denoising rate, thereby achieving the improvement of the echo signal quality of the grounding grid corrosion state under actual environment, and providing a reliable and stable data basis for the observation of the grounding grid corrosion state.

[0068] The technical solution of the present invention is described in detail below in conjunction with various embodiments.

[0069] Example 1

[0070] Please refer to Figure 1 The embodiment of the present invention provides a grounding grid acoustic signal filtering and noise reduction method based on Fourier transform disassembly and parallel acceleration, comprising:

[0071] S0. Obtain the ground grid echo signal and remove the sudden change part through partial filtering.

[0072] Please refer to Figure 2 In this step, partial filtering is implemented based on hardware circuits. Combining the component characteristics of capacitors, inductors and resistors, an RLC circuit based on micro-hardware is constructed to perform constraint range correction on the excessive mutation of the multi-source echo signal.

[0073] S1. Obtain mixed multi-source signals.

[0074] The received multi-source grounding grid echo signal also contains a lot of noise, such as the stray current electromagnetic interference signal of the DC equipment in the high-speed railway station, the grounding grid overvoltage interference signal and the grounding grid echo signal.

[0075] S2. Split the multi-source signal into ground grid echo signal segmentation sub-frequency with different center frequencies.

[0076] By dividing the multi-source sound signals into sub-frequency signals with different center frequencies, a data basis is provided for subsequent noise reduction operations.

[0077] Please refer to Figure 3 , further, the step includes:

[0078] S21, moving the low-pass filter to different center frequencies by a complex modulation method to generate multiple band-pass filter banks;

[0079] S22. Perform multi-phase decomposition on the input multi-source signal, modulate the multi-source signal to baseband, and divide it into a plurality of ground grid echo signal sub-frequency divisions with limited bandwidth and different center frequencies.

[0080] Among them, step S2 is implemented based on a multi-band decomposition module, which divides the ground grid echo signals of multiple sound sources into multiple sub-frequencies with different limited bandwidths with different center frequencies. The ground grid echo signals of multiple sound sources are input, and the ground grid echo signals of multiple sound sources are output as divided sub-frequency groups.

[0081] The specific steps are as follows:

[0082] The design cutoff frequency is The impulse response of the low-pass filter is , the low-pass filter is moved to different center frequencies through complex modulation:

[0083] (1.1)

[0084] (1.2)

[0085] Where w(n) is a window function (such as Kaiser window), each Corresponding center frequency fk, bandwidth .

[0086] For each Perform multiphase decomposition to obtain M branches:

[0087]

[0088] in is the mth polyphase branch of the kth bandpass filter.

[0089] Modulate the input signal x(n) to baseband to generate M sub-band signals:

[0090]

[0091] After modulation, the kth sub-band is moved to the baseband low frequency.

[0092] S3. Perform Fourier transform decomposition and parallel acceleration on the sub-frequency division of the ground grid echo signal to generate parallel component factors of the sub-frequency division of the ground grid echo signal.

[0093] This step addresses the computational complexity of ground grid signal data by leveraging the inherent Fourier transform decomposition operator in the data and combining it with the GPU parallel acceleration module to improve network processing, transforming a single-core computation into a dual-core simultaneous computation. The input data consists of M sub-frequency groups of ground grid echo signals from multiple sound sources, and the output data consists of 2M parallel component factors of these sub-frequency groups.

[0094] Specifically, this step is implemented based on a Fourier transform decomposition parallel acceleration module, which is used to perform at least the following three operations, namely, calculation of two-dimensional continuous Fourier transform, data transposition, and hierarchical decomposition.

[0095] In this embodiment, the ground grid echo signal is divided into sub-frequency segments and Fourier transform is decomposed and accelerated in parallel, including:

[0096] Obtain ground grid echo signal segmentation sub-frequency; convert continuous signals into discrete signals; calculate forward discrete Fourier transform; transpose data and apply operators; Fourier transform decomposition; implement parallel computing.

[0097] Specifically, the detailed implementation of this step is as follows:

[0098] Calculation of two-dimensional continuous Fourier transform: Considering the differential properties of Fourier transform, the sub-frequency division of the ground grid echo signal is calculated in the Fourier domain. The Fourier transform of :

[0099]

[0100] in express The continuous Fourier transform of and is a continuous frequency.

[0101] In the discrete domain, the sub-frequency processing of the ground grid echo signal requires the calculation of the forward discrete Fourier transform:

[0102]

[0103] is a two-dimensional array with 𝑁 A finite square of data points obtained by sampling, and and , and is a discrete frequency, for , .

[0104] Data transposition: Contains the two-dimensional Fourier transform decomposition operator, multiplied by And the data transposition is achieved by converting the discrete domain through backward discrete Fourier transform.

[0105]

[0106] Transformation decomposition: Since the one-dimensional forward and backward transformations cancel each other out, , and Not dependent on and , so (1.6) can be rewritten as:

[0107]

[0108] in and They represent the parallel composition factors of the two ground grid echo signal split sub-frequency, and represent the calculation in the Fourier domain. , where each dimension is calculated independently. In the case of , the one-dimensional kernel involves 、 The length is The forward and backward Fourier transforms of The point-to-point complex multiplication of is sandwiched between them, and a similar definition involves .

[0109] The four nested sums involved in equations (1.6) and (1.7) are reduced to two, thus transforming a single two-dimensional problem into two one-dimensional problems. and The explicit sum makes the Fourier transform decomposition operator dependent on two partial results.

[0110]

[0111] Through the above operations, independent single-core calculations are converted into parallel dual-core calculations (using both CPU and GPU), which naturally achieves the overlap of calculation and communication, thereby improving the calculation speed. Since there is no data dependency between independent one-dimensional calculations, communication transmission of one calculation can be carried out while another calculation is being performed, solving the problem of assigning tasks to different nodes for parallel processing to achieve algorithm efficiency and realize the natural overlap of calculation and communication.

[0112] S4. Perform noise reduction processing on the sub-frequency parallel factors and combine them to obtain a global noise reduction signal.

[0113] The noise reduction processing of the sub-frequency parallel factors in this step includes: performing multi-phase low-pass filtering on each sub-frequency parallel component factor of the ground grid echo signal to filter out high-frequency interference noise; and modulating the filtered sub-frequency parallel component factor of the ground grid echo signal back to the original center frequency.

[0114] Specifically, after the multi-band decomposition module and the parallel channel architecture complete the decomposition and slicing of different center frequencies, the noise reduction processing module uses a previously set modulated low-pass filter to perform filtering processing to filter out interference noise.

[0115] The input data is the parallel component factor of the sub-frequency division of the ground grid echo signal, and the output data is the parallel component factor of the sub-frequency division of the noise-reduced ground grid echo signal.

[0116] Each modulated ground grid echo signal is divided into sub-frequency parallel component factors and polyphase filtering is performed:

[0117]

[0118] in, .

[0119] Combine all branch outputs and modulate back to the original center frequency:

[0120]

[0121] Please refer to Figure 4 In this embodiment, the filtering and noise reduction method further includes:

[0122] S5. Iteratively optimize the global denoised signal through the constrained variation decomposition function.

[0123] Specifically, this step aims to globally refine the initial filtering results using a constrained variation decomposition function to compensate for the limitations of polyphase filtering and parallel architectures. This method suppresses noise while preserving key signal features by constructing an energy function that includes a fidelity term, a smoothness constraint, and a sparsity constraint. The fidelity term ensures the fit between the denoised signal and the original data, preventing excessive distortion. The smoothness constraint utilizes a Fourier transform decomposition operator to suppress high-frequency noise and oscillations, enhancing the signal's time-domain continuity. The sparsity constraint uses the L1 norm to remove isolated noise points while preserving sparse pulse features.

[0124] This step is implemented based on the constrained variation decomposition module, the input is the preliminary denoised signal, and the output is the denoised signal that meets the constraints.

[0125] This step proposes a constrained variation decomposition function to further optimize y(n) and balance the fidelity, smoothness, and sparsity of the ground grid echo signal:

[0126]

[0127] in, is the Fourier transform decomposition operator (second-order difference), λ and μ are weight parameters that control the smoothing and sparsity strength. It is the fidelity item. is a smoothness constraint, Sparse constraints.

[0128] At the same time, the accelerated gradient descent method is used to solve the noise reduction signal y that meets the constraints denoised , the gradient descent iteration formula is shown as (1.14):

[0129]

[0130]

[0131] Where, is the fourth-order difference (the second-order derivative of the smoothing term), and η is the learning rate.

[0132] Example 2

[0133] Please refer to Figure 5 Based on the aforementioned embodiments, the present invention further provides a grounding grid acoustic signal filtering and noise reduction device based on Fourier transform decomposition and parallel acceleration (hereinafter referred to as the filtering and noise reduction device), comprising:

[0134] The multi-band decomposition module 10 is configured to divide the multi-source signal into ground grid echo signal segmentation sub-frequency with different center frequencies, providing a data basis for subsequent noise reduction operations.

[0135] Please refer to Figure 6 Specifically, the multi-band decomposition module includes:

[0136] The RLC circuit 101 has a circuit structure as follows: Figure 2 As shown, it is used to perform constraint range correction on the excessive mutation of multi-source signals;

[0137] The complex modulation unit 102 is used to move the low-pass filter to different center frequencies to generate a band-pass filter bank.

[0138] Furthermore, the filtering and noise reduction device further includes:

[0139] The Fourier transform decomposition parallel acceleration module 20 is configured to perform two-dimensional Fourier transform decomposition and GPU parallel acceleration calculation on the ground grid echo signal divided into sub-frequency segments to generate parallel component factors of the ground grid echo signal divided into sub-frequency segments.

[0140] The noise reduction processing module 30 is configured to perform multi-phase low-pass filtering and signal merging on the sub-frequency parallel component factors of the ground grid echo signal to obtain a global noise reduction signal.

[0141] Furthermore, the grounding grid acoustic signal filtering and noise reduction device based on Fourier transform decomposition and parallel acceleration provided in this embodiment further includes:

[0142] The constrained variation decomposition module 40 is configured to iteratively optimize the global noise reduction signal through an energy function.

[0143] Please refer to Figure 7 , wherein the Fourier transform disassembly parallel acceleration module 20 includes:

[0144] GPU computing unit 201, for performing two one-dimensional Fourier transform and complex multiplication operations in parallel;

[0145] The data transposition unit 202 is configured to separate the horizontal and vertical frequency domain components.

[0146] It is understandable that the steps performed by the filtering and noise reduction device are the same as those in the aforementioned embodiment 1, and will not be described in detail here.

[0147] Finally, it should be noted that although the above embodiments have been described in the specification and drawings of this application, this does not limit the scope of protection of the invention of this application. All technical solutions generated by replacing or modifying equivalent structures or equivalent processes based on the essential concepts of this application using the contents recorded in the specification and drawings of this application, as well as directly or indirectly implementing the technical solutions of the above embodiments in other related technical fields, are included in the scope of protection of the invention of this application.

Claims

1. A grounding grid acoustic signal filtering and noise reduction method based on Fourier transform disassembly and parallel acceleration, characterized in that: include: Obtain mixed multi-source signals; Splitting the multi-source signal into ground grid echo signal split sub-frequency components with different center frequencies; Performing Fourier transform decomposition and parallel acceleration on the ground grid echo signal divided sub-frequency to generate parallel component factors of the ground grid echo signal divided sub-frequency; Noise reduction processing is performed on the sub-frequency parallel factors, and a global noise reduction signal is obtained after merging.

2. The grounding grid acoustic signal filtering and noise reduction method based on Fourier transform disassembly and parallel acceleration according to claim 1 is characterized in that: The step of dividing the multi-source signal into ground grid echo signal division sub-frequency components with different center frequencies includes: The low-pass filter is moved to different center frequencies through complex modulation method to generate multiple band-pass filter banks; The input multi-source signal is subjected to multi-phase decomposition, modulated to a baseband, and divided into a plurality of ground grid echo signal sub-frequency divisions with limited bandwidth and different center frequencies.

3. The grounding grid acoustic signal filtering and noise reduction method based on Fourier transform disassembly and parallel acceleration according to claim 1 is characterized in that: The parallel acceleration of Fourier transform decomposition of the ground grid echo signal into sub-frequency segments includes: Get the ground grid echo signal split sub-frequency , calculate the ground grid echo signal segmentation sub-frequency in Fourier domain Fourier transform of ;in express The continuous Fourier transform of and is the continuous frequency; right Sampling is performed to obtain a two-dimensional array , realizing the conversion from continuous domain to discrete domain; the two-dimensional array By having A finite square of data points Sampling is performed, in which and , and is a discrete frequency, for , ; right Calculate the forward discrete Fourier transform to obtain its frequency domain representation; Using the two-dimensional Fourier transform to decompose the operator, Multiply , and converted back to the discrete domain via backward discrete Fourier transform; Will Disassembled into , transform the two-dimensional problem into two one-dimensional problems, and get and Two parallel component factors.

4. The method for filtering and denoising grounding grid acoustic signals based on Fourier transform disassembly and parallel acceleration according to claim 1 is characterized in that: The performing noise reduction processing on the sub-frequency parallel factors includes: Each grounding grid echo signal is divided into sub-frequency parallel component factors and multi-phase low-pass filtering is performed to filter out high-frequency interference noise; The filtered ground grid echo signal is divided into sub-frequency components and modulated back to the original center frequency in parallel.

5. The grounding grid acoustic signal filtering and noise reduction method based on Fourier transform disassembly and parallel acceleration according to claim 1 is characterized in that: Also includes: Based on the following formula, the global noise reduction signal is iteratively optimized through the constrained variation decomposition function: ； in, is the Fourier transform decomposition operator (second-order difference), λ and μ are weight parameters that control the smoothing and sparsity strength; For authenticity, is a smoothness constraint, is a sparse constraint.

6. The method for filtering and denoising ground grid acoustic signals based on Fourier transform disassembly and parallel acceleration according to claim 5 is characterized in that: The iterative optimization includes: The accelerated gradient descent method is used to solve the denoised signal ydenoised that satisfies the constraints based on the following gradient descent formula: ； ； Where, is the fourth-order difference (the second-order derivative of the smoothing term), and η is the learning rate.

7. A grounding grid acoustic signal filtering and noise reduction device based on Fourier transform disassembly and parallel acceleration, characterized in that: include: a multi-band decomposition module configured to split the multi-source signal into ground grid echo signal segmentation sub-frequency components with different center frequencies; A Fourier transform decomposition parallel acceleration module is configured to perform two-dimensional Fourier transform decomposition and GPU parallel acceleration calculation on the sub-frequency segmentation of the ground grid echo signal to generate parallel component factors of the sub-frequency segmentation of the ground grid echo signal; The noise reduction processing module is configured to perform multiphase low-pass filtering and signal merging on the parallel component factors of the ground grid echo signal segmentation sub-frequency to obtain a global noise reduction signal.

8. The grounding grid acoustic signal filtering and noise reduction device based on Fourier transform disassembly and parallel acceleration according to claim 7 is characterized in that: Also includes: The constrained variation decomposition module is configured to iteratively optimize the global denoised signal through an energy function.

9. The grounding grid acoustic signal filtering and noise reduction device based on Fourier transform disassembly and parallel acceleration according to claim 7 is characterized in that: The multi-band decomposition module includes: RLC circuit is used to correct the excessive mutation of multi-source signals within the constraint range; The complex modulation unit is used to move the low-pass filter to different center frequencies to generate a band-pass filter bank.

10. The grounding grid acoustic signal filtering and noise reduction device based on Fourier transform disassembly and parallel acceleration according to claim 7 is characterized in that: The Fourier transform disassembly parallel acceleration module includes: GPU computing unit for performing two one-dimensional Fourier transform and complex multiplication operations in parallel; The data transposition unit is used to separate the frequency domain components in the horizontal and vertical directions.

Images