FID signal tuning method, device and equipment

Abstract

The invention provides an FID signal tuning method, device and equipment, and relates to the technical field of information processing, and the method comprises the steps: carrying out the matrix segmentation of a to-be-processed FID signal, and obtaining a corresponding fourth-order tensor; decomposing the fourth-order tensor by using an Nth-SVD method to obtain a denoised FID (Flame Ionization Detector) signal; an HOC-Prony algorithm is adopted to process the FID signal, and the amplitude, the frequency and the attenuation factor of a fourth-order mixed cumulant diagonal slice of the FID signal are determined; and determining the tuning frequency of the FID signal to be processed according to the frequency band range of the FID signal, the amplitude, the frequency and the attenuation factor. According to the method, the high-order tensor is constructed by using the matrix segmentation method, the signal component and the noise component are separated in the tensor space, the signal component is effectively extracted by using the tensor decomposition method, the high-order cumulant is introduced, and the noise immunity of the Prony model is enhanced, so that the accuracy and efficiency of signal tuning frequency estimation are improved.

Description

Technical Field

[0001] The present invention relates to the field of information processing technology, and in particular to a FID signal tuning method, device and equipment. Background Art

[0002] The Earth's magnetic field is an important information resource with wide application value in many fields, such as archeology, oil and gas pipeline detection, geological exploration, etc. The proton precession magnetometer is an instrument used to measure the magnitude of the Earth's magnetic field. The free induction decay (FID) signal induced by the sensor is proportional to the magnetic field to be measured. The higher the signal frequency estimation accuracy, the higher the magnetic field measurement accuracy. However, due to the interference of Gaussian random noise in the environment, the initial signal-to-noise ratio of the FID signal is low. Therefore, the magnetometer needs to use a resonant capacitor to tune the sensor to suppress noise and improve the frequency estimation accuracy.

[0003] However, the traditional tuning method is only applicable to noise-free or weak noise environments, and the tuning method of commercial magnetometers is inefficient and has low tuning accuracy, and is prone to "detuning". With the continuous expansion of human activities, complex electromagnetic interference caused by various electrical equipment has emerged. The current tuning method can no longer meet the current magnetometer detection needs, and it is urgent to study the precise tuning method of proton precession magnetometers in strong noise environments. Summary of the invention

[0004] The purpose of the present invention is to solve the problem that the current tuning method has low tuning efficiency and tuning accuracy and cannot meet the current magnetometer detection needs. The present invention provides a FID signal tuning method, device and equipment.

[0005] The technical solution of the embodiment of the present application is implemented as follows: A first aspect of an embodiment of the present application provides a FID signal tuning method, comprising: Perform matrix segmentation on the FID signal to be processed to obtain the corresponding fourth-order tensor; Decomposing the fourth-order tensor by using the Nth-SVD method to obtain a denoised FID signal; The FID signal is processed by using the HOC-Prony algorithm to determine the amplitude, frequency and attenuation factor of the diagonal slice of the fourth-order mixed cumulant of the FID signal; The tuning frequency of the FID signal to be processed is determined according to the frequency band range of the FID signal, the amplitude, the frequency and the attenuation factor.

[0006] Optionally, before performing matrix segmentation on the FID signal to be processed to obtain a corresponding fourth-order tensor, the method further includes: Multiple FID signals to be processed are acquired using an equal delay method.

[0007] Optionally, performing matrix segmentation on the FID signal to be processed to obtain a corresponding fourth-order tensor includes: The FID signal x=[s1, s2, s3, …, sN] to be processed with a length of N is divided into K segments, and the dimension of each segment is I×J, where , the FID signal x to be processed can be expressed as: ; In the formula, the expressions of each segment matrix are as follows: .

[0008] Optionally, the decomposing the fourth-order tensor by using the Nth-SVD method to obtain a denoised FID signal includes: Decomposing the fourth-order tensor to obtain an initial core tensor and multiple accompanying unitary matrices; in the initial core tensor, the Frobenius norm of the sub-tensor satisfies the order, obtaining the difference between adjacent sub-tensors, selecting the position with significant difference as the threshold, the sub-tensor greater than the threshold is the signal component, and the sub-tensor less than the threshold is the noise component; filtering out the noise component in the initial core tensor by using a differential method, and determining the signal component as a target core tensor; The denoised target fourth-order tensor is determined based on the target core tensor and the adjoint unitary matrix, and the denoised FID signal is obtained by using an inverse matrix partitioning transform.

[0009] Optionally, the adopting of the HOC-Prony algorithm to process the FID signal to determine the amplitude, frequency and attenuation factor of the diagonal slice of the fourth-order mixed cumulant of the FID signal includes: Acquire a fourth-order mixed cumulant slice of the FID signal; The fourth-order mixed cumulant slices are fitted using a linear combination of exponential functions to obtain a Prony fitting equation; The Prony fitting equation is solved by the least square method to determine the amplitude, frequency and attenuation factor of the diagonal slice of the fourth-order mixed cumulant of the FID signal.

[0010] Optionally, the fourth-order mixed cumulant slices are fitted using a linear combination of exponential functions to obtain a Prony fitting equation, including: Determine the mixing cumulative amount using the following formula : ; In the formula, represents the cumulant, E[·] represents the mathematical expectation, and τ1, τ2, and τ3 represent any lag values ​​of the cumulant. When τ1=τ2=τ3=τ, the fourth-order mixed cumulant slice for: ; A set of linear combinations of p exponential functions is used The specific expression is as follows: ; Where N is the number of signal sampling points, and the expressions of bk and zk are as follows: ; In the formula, is the amplitude, is the phase, is the frequency, is the attenuation factor, is the sampling interval.

[0011] Optionally, determining the tuning frequency of the FID signal to be processed according to the frequency band range, the amplitude, the frequency and the attenuation factor of the FID signal includes: Screening is performed according to the frequency band range of the FID signal to determine a target frequency band range; Based on the amplitude and attenuation factor corresponding to each signal frequency within the target frequency band, a screening factor is obtained, and the signal frequency corresponding to the maximum screening factor is determined as the tuning frequency of the FID signal to be processed.

[0012] A second aspect of an embodiment of the present application provides a FID signal tuning device, including: a matrix segmentation module, a tensor decomposition module, a processing module and a determination module, wherein: The matrix segmentation module is configured to perform matrix segmentation on the FID signal to be processed to obtain a corresponding fourth-order tensor; The tensor decomposition module is configured to decompose the fourth-order tensor using an Nth-SVD method to obtain a denoised FID signal; The processing module is configured to process the FID signal using a HOC-Prony algorithm to determine the amplitude, frequency and attenuation factor of a diagonal slice of a fourth-order mixed cumulant of the FID signal; The determination module is configured to determine the tuning frequency of the to-be-processed FID signal according to the frequency band range, the amplitude, the frequency and the attenuation factor of the FID signal.

[0013] A third aspect of an embodiment of the present application provides an electronic device, including a processor and a memory; the memory stores a computer program, wherein the computer program implements the FID signal tuning method described in the first aspect when executed by the processor.

[0014] Compared with the prior art, the technical solution provided by this application has the following beneficial effects: The present invention provides a FID signal tuning method and device, which performs matrix segmentation on a to-be-processed FID signal to obtain a corresponding fourth-order tensor, decomposes the fourth-order tensor using an Nth-SVD method to obtain a denoised FID signal, processes the FID signal using an HOC-Prony algorithm, determines the amplitude, frequency and attenuation factor of a diagonal slice of a fourth-order mixed cumulant of the FID signal, and then determines the tuning frequency of the to-be-processed FID signal according to the frequency band range, amplitude, frequency and attenuation factor of the FID signal. The present invention constructs a high-order tensor using a matrix segmentation method, separates a signal component and a noise component in a tensor space, and effectively extracts the signal component using a tensor decomposition method, introduces a high-order cumulant, enhances the noise resistance of the Prony model, and thus improves the accuracy and efficiency of signal tuning frequency estimation. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A schematic diagram of a flow chart of a FID signal tuning method provided in an embodiment of the present application; Figure 2 This is the curve of the root mean square error of the tuning frequency deviation changing with Gaussian random noise of different intensities under the traditional tuning method; Figure 3 The root mean square error of the tuning frequency deviation changes with different intensities of Gaussian random noise under the method of the present invention. Figure 4 A schematic diagram of the structure of the FID signal tuning device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0016] Below, embodiments of the present application will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present application. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present application. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present application.

[0017] The terms used herein are only for describing specific embodiments and are not intended to limit the present application. The terms "include", "comprising", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components. All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0018] Some block diagrams and / or flow charts are shown in the accompanying drawings. It should be understood that some blocks or combinations thereof in the block diagrams and / or flow charts may be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that these instructions, when executed by the processor, may create a device for implementing the functions / operations described in these block diagrams and / or flow charts.

[0019] In some embodiments, see Figure 1 , Figure 1 A flow chart of a FID signal tuning method provided in an embodiment of the present application; the FID signal tuning method provided in an embodiment of the present application includes: S110, performing matrix segmentation on the FID signal to be processed to obtain a corresponding fourth-order tensor.

[0020] In a magnetometer, the free induced decay (FID) signal is a sinusoidal wave with a constant frequency and an amplitude that decays with time, generated in the induction circuit by a polarized proton system performing precession motion under the influence of the Earth's magnetic field.

[0021] The fourth-order tensor, i.e., the tensor representing the multi-channel FID signal after matrix segmentation, has four dimensions I, J, K, and Q. Among them, Q represents the number of channels in the signal acquisition process, and I, J, and K satisfy the following conditional relationship: ; Where N represents the signal length, m is a natural number, is a set of natural numbers.

[0022] In some embodiments, at S110, before performing matrix segmentation on the FID signal to be processed to obtain a corresponding fourth-order tensor, the FID signal tuning method further includes: Multiple FID signals to be processed are acquired using an equal delay method.

[0023] Here, after each RF pulse is transmitted, a fixed delay time (such as a few milliseconds) is waited to avoid the influence of dead time on signal acquisition, and then the FID signal is collected. During the entire experiment, multi-channel synchronous acquisition of FID signals is used to ensure that all FID signals have consistency in acquisition time. The specific delay time can be measured based on experience or experiment.

[0024] In some embodiments, S110, matrix segmentation is performed on the FID signal to be processed to obtain a corresponding fourth-order tensor, including: The FID signal x=[s1, s2, s3, …, sN] to be processed with a length of N is divided into K segments, and the dimension of each segment is I×J, where , the FID signal x to be processed can be expressed as: ; In the formula, the expressions of each segment matrix are as follows: .

[0025] S120, decomposing the fourth-order tensor using an Nth-SVD method to obtain a denoised FID signal.

[0026] The Nth-SVD method, that is, the singular value decomposition method, can specifically include Tucker decomposition and CANDECOMP / PARAFAC (CP) decomposition. Among them, Tucker decomposition is one of the most commonly used high-order tensor decomposition methods, which can decompose an N-order tensor into the product of a core tensor and N factor matrices, similar to the SVD of a matrix, but extended to multiple dimensions. CANDECOMP / PARAFAC (CP) decomposition can decompose a tensor into the sum of a series of rank tensors, where each rank tensor is the outer product of multiple vectors.

[0027] In some embodiments, S120, using the Nth-SVD method to decompose the fourth-order tensor to obtain a denoised FID signal includes: Decompose the fourth-order tensor to obtain an initial core tensor and multiple accompanying unitary matrices; in the initial core tensor, the Frobenius norm of the sub-tensor satisfies the order, obtain the difference between adjacent sub-tensors, select the position with significant difference as the threshold, the sub-tensor greater than the threshold is the signal component, and the sub-tensor less than the threshold is the noise component; The noise component in the initial core tensor is filtered out by differential method, and the signal component is determined as the target core tensor; The denoised target fourth-order tensor is determined based on the target core tensor and the adjoint unitary matrix, and the denoised FID signal is obtained by using the inverse matrix partitioning transform.

[0028] S130, using the HOC-Prony algorithm to process the FID signal, and determining the amplitude, frequency and attenuation factor of the diagonal slice of the fourth-order mixed cumulant of the FID signal.

[0029] The Prony algorithm is a mathematical method for analyzing and identifying oscillatory components in equally spaced sampled data. The data is fitted by constructing a linear combination of a series of complex exponential decay functions to extract the amplitude, phase, frequency and attenuation coefficient of each component. Here, the high-order statistics HOC and the Prony algorithm are combined to process the FID signal.

[0030] In some embodiments, S130, the FID signal is processed using the HOC-Prony algorithm to determine the amplitude, frequency and attenuation factor of the diagonal slice of the fourth-order mixed cumulant of the FID signal, including: Obtain a fourth-order mixed cumulant slice of the FID signal; The Prony fitting equation is obtained by fitting the fourth-order mixed cumulant slices using the linear combination of exponential functions. The Prony fitting equation is solved by the least square method to determine the amplitude, frequency and attenuation factor of the diagonal slice of the fourth-order mixed cumulant of the FID signal.

[0031] In some embodiments, the fourth-order mixed cumulant slices are fitted using a linear combination of exponential functions to obtain a Prony fitting equation, including: Determine the mixing cumulative volume using the following formula: ; In the formula, represents the cumulant, E[·] represents the mathematical expectation, and τ1, τ2, and τ3 represent any lag values ​​of the cumulant. When τ1=τ2=τ3=τ, the fourth-order mixed cumulant slice for: ; A set of linear combinations of p exponential functions is used The specific expression is as follows: ; Where N is the number of signal sampling points, and the expressions of bk and zk are as follows: ; In the formula, is the amplitude, is the phase, is the frequency, is the attenuation factor, is the sampling interval.

[0032] In an optional embodiment, the key to fitting a given signal by the Prony model is to find a homogeneous solution to a linear difference equation with constant coefficients. The difference equation is as follows: ; In the solution process, the second-order derivative of the singular value is introduced in advance to obtain the model order p, and the above difference equation is used to solve it, and then The characteristic equation with parameters is obtained ,Right now: ; Solve the Prony fitting equation using the least squares method: ; Available, .in accordance with , and , , , The relationship between them can be obtained: .

[0033] S140, determining a tuning frequency of the FID signal to be processed according to the frequency band range, amplitude, frequency and attenuation factor of the FID signal.

[0034] In this embodiment, the frequency band of the FID signal is used for coarse screening, and the amplitude, frequency and attenuation factor are used for fine screening.

[0035] In some embodiments, S140, determining the tuning frequency of the FID signal to be processed according to the frequency band range, amplitude, frequency and attenuation factor of the FID signal includes: Screening is performed according to the frequency band range of the FID signal to determine the target frequency band range; Based on the amplitude and attenuation factor corresponding to each signal frequency within the target frequency band, a screening factor is obtained, and the signal frequency corresponding to the maximum screening factor is determined as the tuning frequency of the FID signal to be processed.

[0036] In a specific embodiment, the effect of the FID signal tuning method proposed by the present invention is verified and compared with the traditional tuning method. Table 1 shows the root mean square error of the FID signal tuning frequency deviation after tuning by different methods under different strengths of Gaussian random noise interference.

[0037]

[0038] Table 1 Figure 2 This is the curve of the root mean square error of the tuning frequency deviation changing with Gaussian random noise of different intensities under the traditional tuning method; Figure 3 This is a curve showing the change of the root mean square error of the tuning frequency deviation with Gaussian random noise of different intensities under the method proposed in the present invention.

[0039] According to Table 1 and Figure 2 , Figure 3 It can be seen that in a strong noise environment, that is, when the initial signal-to-noise ratio of the FID signal is less than -20dB, the tuning frequency deviation of the AC-FFT and ST-SVD methods exceeds 1000 Hz, and the accurate tuning frequency cannot be obtained. When the initial signal-to-noise ratio of the FID signal is greater than -10 dB, AC-FFT and ST-SVD can both accurately extract the tuning frequency, and the tuning frequency deviation is within 10Hz, but the tuning frequency deviation obtained by the method proposed in the present invention is smaller, and the tuning frequency deviation is within 8Hz. From the above analysis, it can be seen that the method proposed in the present invention can accurately obtain the signal tuning frequency under Gaussian noise interference of different intensities.

[0040] The embodiment of the present application performs matrix segmentation on the FID signal to be processed to obtain the corresponding fourth-order tensor, decomposes the fourth-order tensor using the Nth-SVD method to obtain the denoised FID signal, processes the FID signal using the HOC-Prony algorithm, determines the amplitude, frequency and attenuation factor of the diagonal slice of the fourth-order mixed cumulant of the FID signal, and then determines the tuning frequency of the FID signal to be processed according to the frequency band range, amplitude, frequency and attenuation factor of the FID signal. The present invention constructs a high-order tensor using a matrix segmentation method, separates signal components and noise components in the tensor space, and effectively extracts signal components using a tensor decomposition method, introduces high-order cumulants, enhances the noise resistance of the Prony model, and thus improves the accuracy and efficiency of signal tuning frequency estimation.

[0041] In some embodiments, see Figure 4 , Figure 4 The structural diagram of a FID signal tuning device provided in an embodiment of the present application is shown in FIG. The FID signal tuning device 400 provided in an embodiment of the present application includes: a matrix segmentation module 410, a tensor decomposition module 420, a processing module 430 and a determination module 440, wherein: The matrix segmentation module 410 is configured to perform matrix segmentation on the FID signal to be processed to obtain a corresponding fourth-order tensor.

[0042] The tensor decomposition module 420 is configured to decompose the fourth-order tensor using the Nth-SVD method to obtain a denoised FID signal.

[0043] The processing module 430 is configured to process the FID signal using the HOC-Prony algorithm to determine the amplitude, frequency and attenuation factor of the diagonal slice of the fourth-order mixed cumulant of the FID signal.

[0044] The determination module 440 is configured to determine the tuning frequency of the FID signal to be processed according to the frequency band range, amplitude, frequency and attenuation factor of the FID signal.

[0045] In some embodiments, the FID signal tuning device 400 further includes an acquisition module, and the acquisition module is specifically configured as follows: Multiple FID signals to be processed are acquired using an equal delay method.

[0046] In some embodiments, the matrix segmentation module 410 is specifically configured as follows: The FID signal x=[s1, s2, s3, …, sN] to be processed with a length of N is divided into K segments, and the dimension of each segment is I×J, where , the FID signal x to be processed can be expressed as: ; In the formula, the expressions of each segment matrix are as follows: .

[0047] In some embodiments, the tensor decomposition module 420 is specifically configured as follows: Decompose the fourth-order tensor to obtain an initial core tensor and multiple accompanying unitary matrices; in the initial core tensor, the Frobenius norm of the sub-tensor satisfies the order, obtain the difference between adjacent sub-tensors, select the position with significant difference as the threshold, the sub-tensor greater than the threshold is the signal component, and the sub-tensor less than the threshold is the noise component; The noise component in the initial core tensor is filtered out by differential method, and the signal component is determined as the target core tensor; The denoised target fourth-order tensor is determined based on the target core tensor and the adjoint unitary matrix, and the denoised FID signal is obtained by using the inverse matrix partitioning transform.

[0048] In some embodiments, the processing module 430 is specifically configured as follows: Obtain a fourth-order mixed cumulant slice of the FID signal; The Prony fitting equation is obtained by fitting the fourth-order mixed cumulant slices using the linear combination of exponential functions. The Prony fitting equation is solved by the least square method to determine the amplitude, frequency and attenuation factor of the diagonal slice of the fourth-order mixed cumulant of the FID signal.

[0049] In some embodiments, the processing module 430 is further configured to: Determine the mixing cumulative volume using the following formula: ; In the formula, represents the cumulant, E[·] represents the mathematical expectation, and τ1, τ2, and τ3 represent any lag values ​​of the cumulant. When τ1=τ2=τ3=τ, the fourth-order mixed cumulant slice for: ; A set of linear combinations of p exponential functions is used The specific expression is as follows: ; Where N is the number of signal sampling points, and the expressions of bk and zk are as follows: ; In the formula, is the amplitude, is the phase, is the frequency, is the attenuation factor, is the sampling interval.

[0050] In some embodiments, the determination module 440 is specifically configured to: Screening is performed according to the frequency band range of the FID signal to determine the target frequency band range; Based on the amplitude and attenuation factor corresponding to each signal frequency within the target frequency band, a screening factor is obtained, and the signal frequency corresponding to the maximum screening factor is determined as the tuning frequency of the FID signal to be processed.

[0051] The FID signal tuning device provided in the embodiment of the present application can implement each process in the embodiment corresponding to the above-mentioned FID signal tuning method, and will not be described again here to avoid repetition.

[0052] It should be noted that the FID signal tuning device provided in the embodiment of the present application and the FID signal tuning method provided in the embodiment of the present application are based on the same application concept, so the specific implementation of this embodiment can refer to the implementation of the aforementioned FID signal tuning method, and the repeated parts will not be repeated.

[0053] Those skilled in the art will appreciate that the features described in the various embodiments and / or claims of the present application may be combined and / or combined in a variety of ways, even if such combinations or combinations are not explicitly described in the present application. In particular, without departing from the spirit and teachings of the present application, the features described in the various embodiments and / or claims of the present application may be combined and / or combined in a variety of ways. All of these combinations and / or combinations fall within the scope of the present application. Therefore, the scope of the present application should not be limited to the above-described embodiments, but should be determined not only by the appended claims, but also by the equivalents of the appended claims.

Claims

1. A FID signal tuning method, characterized in that: include: Perform matrix segmentation on the FID signal to be processed to obtain the corresponding fourth-order tensor; Decomposing the fourth-order tensor by using the Nth-SVD method to obtain a denoised FID signal; The FID signal is processed by using the HOC-Prony algorithm to determine the amplitude, frequency and attenuation factor of the diagonal slice of the fourth-order mixed cumulant of the FID signal; The tuning frequency of the FID signal to be processed is determined according to the frequency band range of the FID signal, the amplitude, the frequency and the attenuation factor.

2. The FID signal tuning method according to claim 1, characterized in that: Before performing matrix segmentation on the FID signal to be processed to obtain a corresponding fourth-order tensor, the method further includes: Multiple FID signals to be processed are acquired using an equal delay method.

3. The FID signal tuning method according to claim 1, characterized in that: The matrix segmentation of the FID signal to be processed is performed to obtain a corresponding fourth-order tensor, including: The FID signal x=[s1, s2, s3, …, sN] to be processed with a length of N is divided into K segments, and the dimension of each segment is I×J, where , the FID signal x to be processed can be expressed as: ； In the formula, the expressions of each segment matrix are as follows: 。 4. The FID signal tuning method according to claim 1, characterized in that: Decomposing the fourth-order tensor by using the Nth-SVD method to obtain a denoised FID signal includes: Decomposing the fourth-order tensor to obtain an initial core tensor and multiple accompanying unitary matrices; in the initial core tensor, the Frobenius norm of the sub-tensor satisfies the order, obtaining the difference between adjacent sub-tensors, selecting the position with significant difference as the threshold, the sub-tensor greater than the threshold is the signal component, and the sub-tensor less than the threshold is the noise component; filtering out the noise component in the initial core tensor by using a differential method, and determining the signal component as a target core tensor; The denoised target fourth-order tensor is determined based on the target core tensor and the adjoint unitary matrix, and the denoised FID signal is obtained by using an inverse matrix partitioning transform.

5. The FID signal tuning method according to claim 1, characterized in that: The HOC-Prony algorithm is used to process the FID signal to determine the amplitude, frequency and attenuation factor of the diagonal slice of the fourth-order mixed cumulant of the FID signal, including: Acquire a fourth-order mixed cumulant slice of the FID signal; The fourth-order mixed cumulant slices are fitted using a linear combination of exponential functions to obtain a Prony fitting equation; The Prony fitting equation is solved by the least square method to determine the amplitude, frequency and attenuation factor of the diagonal slice of the fourth-order mixed cumulant of the FID signal.

6. The FID signal tuning method according to claim 5, characterized in that: The fourth-order mixed cumulant slices are fitted by using a linear combination of exponential functions to obtain a Prony fitting equation, including: Determine the mixing cumulative amount using the following formula : ； In the formula, represents the cumulant, E[·] represents the mathematical expectation, and τ1, τ2, and τ3 represent any lag values ​​of the cumulant. When τ1=τ2=τ3=τ, the fourth-order mixed cumulant slice for: ； A set of linear combinations of p exponential functions is used The specific expression is as follows: ； Where N is the number of signal sampling points, and the expressions of bk and zk are as follows: ； In the formula, is the amplitude, is the phase, is the frequency, is the attenuation factor, is the sampling interval.

7. The FID signal tuning method according to claim 1, characterized in that: The step of determining the tuning frequency of the FID signal to be processed according to the frequency band range, the amplitude, the frequency and the attenuation factor of the FID signal comprises: Screening is performed according to the frequency band range of the FID signal to determine a target frequency band range; Based on the amplitude and attenuation factor corresponding to each signal frequency within the target frequency band, a screening factor is obtained, and the signal frequency corresponding to the maximum screening factor is determined as the tuning frequency of the FID signal to be processed.

8. A FID signal tuning device, characterized in that: include: Matrix segmentation module, tensor decomposition module, processing module and determination module, wherein, The matrix segmentation module is configured to perform matrix segmentation on the FID signal to be processed to obtain a corresponding fourth-order tensor; The tensor decomposition module is configured to decompose the fourth-order tensor using an Nth-SVD method to obtain a denoised FID signal; The processing module is configured to process the FID signal using a HOC-Prony algorithm to determine the amplitude, frequency and attenuation factor of a diagonal slice of a fourth-order mixed cumulant of the FID signal; The determination module is configured to determine the tuning frequency of the to-be-processed FID signal according to the frequency band range, the amplitude, the frequency and the attenuation factor of the FID signal.

9. An electronic device comprising a processor and a memory; the memory stores a computer program, wherein: When the computer program is executed by the processor, the computer program implements the FID signal tuning method according to any one of claims 1 to 7.

Images