A method and system for extracting ground nuclear magnetic resonance signal parameters

Abstract

The invention discloses a method and system for extracting ground nuclear magnetic resonance signal parameters, and relates to the technical field of nuclear magnetic resonance data processing. The method comprises: acquiring ground nuclear magnetic resonance data; constructing a nuclear magnetic signal model according to the ground nuclear magnetic resonance data; constructing an output vector and a random noise vector according to the ground nuclear magnetic resonance data; constructing a state vector according to an SNMR signal submodel and a power frequency harmonic noise submodel; constructing a state space equation according to the output vector, the state vector and the random noise vector; solving the state space equation to obtain a solution of a system matrix; determining the relaxation time of the SNMR signal and the Larmor frequency of the SNMR signal according to an eigenvalue with the smallest absolute value of the real part in the solution of the system matrix. The invention can improve the accuracy of parameter extraction and reduce the amount of calculation.

Description

Technical Field

[0001] The present invention relates to the technical field of nuclear magnetic resonance data processing, and in particular to a method and system for extracting ground nuclear magnetic resonance signal parameters. Background Art

[0002] The nuclear magnetic resonance sounding method for finding water works by emitting a pulse signal underground, causing the hydrogen protons in the underground aquifer to produce nuclear magnetic resonance. By collecting the echoes reflected by the hydrogen protons and estimating key parameters of the signal, the water storage capacity of the underground aquifer, depth from the ground, groundwater conductivity and average porosity of underground pores can be determined.

[0003] The nuclear magnetic resonance water detection method has the advantages of high resolution, rich hydrological information, low detection cost and high efficiency, but there are also some problems to be solved. The surface nuclear magnetic resonance (SNMR) signal collected at the detection site is usually very weak, with an amplitude of only a few nanovolts to hundreds of nanovolts, and there are often a variety of noises in the acquisition environment that affect the signal quality. The extremely low signal-to-noise ratio makes it very difficult to extract parameters from the SNMR signal. Power frequency harmonic noise and environmental random noise are the most common noises mixed in the collected SNMR data.

[0004] At present, the commonly used methods for extracting relaxation time and Larmor frequency parameters of SNMR signals can be divided into three categories:

[0005] 1. The strategy of first eliminating noise to improve the signal-to-noise ratio and then performing curve fitting to obtain the ground NMR signal parameters. This method needs to first target the noise characteristics and use various denoising methods, such as harmonic modeling to estimate the power frequency harmonic noise first, and multiple superposition method to filter out random noise first. This method is not only affected by the accuracy of curve fitting, but also by the denoising method, and has limited accuracy.

[0006] 2. ESPRIT and other correlation methods or high-order statistics methods. The amplitude of the SNMR signal varies with time. Directly performing correlation methods or high-order statistics calculations on it will degrade the performance and affect the estimation accuracy.

[0007] 3. Maximum likelihood method. The maximum likelihood method has high estimation accuracy and can process time-varying signals, but the amount of calculation is too large, which is not conducive to practical application.

[0008] In summary, the existing SNMR signal parameter extraction method is used to extract the relaxation time and Larmor frequency of the SNMR signal, but there are problems such as low parameter extraction accuracy and excessive calculation. Summary of the invention

[0009] The purpose of the present invention is to provide a method and system for extracting parameters of ground nuclear magnetic resonance signals, so as to improve the accuracy of parameter extraction and reduce the amount of calculation.

[0010] To achieve the above object, the present invention provides the following solutions:

[0011] A method for extracting ground nuclear magnetic resonance signal parameters, the method comprising:

[0012] Acquire ground nuclear magnetic resonance data; the ground nuclear magnetic resonance data includes a plurality of sampling points; each of the sampling points includes SNMR signals, power frequency harmonic noise and random noise;

[0013] Constructing a nuclear magnetic resonance signal model according to the ground nuclear magnetic resonance data; the nuclear magnetic resonance signal model includes a SNMR signal sub-model, a power frequency harmonic noise sub-model and a random noise sub-model;

[0014] constructing an output vector and a random noise vector according to the ground nuclear magnetic resonance data;

[0015] Constructing a state vector according to the SNMR signal sub-model and the power frequency harmonic noise sub-model;

[0016] constructing a state space equation according to the output vector, the state vector and the random noise vector;

[0017] Solving the state space equation to obtain a solution of a system matrix; the solution of the system matrix includes a plurality of eigenvalues;

[0018] The relaxation time of the SNMR signal and the Larmor frequency of the SNMR signal are determined according to an eigenvalue with the smallest absolute value of the real part in the solution of the system matrix.

[0019] Optionally, constructing a nuclear magnetic resonance signal model according to the ground nuclear magnetic resonance data specifically includes:

[0020] Constructing a SNMR signal sub-model according to the SNMR signal;

[0021] Constructing a power frequency harmonic noise sub-model according to the power frequency harmonic noise;

[0022] A random noise sub-model is constructed according to the random noise.

[0023] Optionally, the nuclear magnetic signal model is expressed as:

[0024]

[0025] Wherein, y(n) represents the ground NMR data, n=1, 2, ..., N, N represents the total number of the sampling points, and n represents the sequence number of the sampling point. represents the SNMR signal submodel, represents the power frequency harmonic noise sub-model, w(n) represents the random noise sub-model, E 0 represents the initial amplitude of the SNMR signal, represents the relaxation time of the SNMR signal, j represents the imaginary number sign, and f L represents the Larmor frequency of the SNMR signal, A represents the initial phase of the SNMR signal, B represents the total number of power frequency harmonic components, b represents the serial number of the power frequency harmonic component, b=1,2,...,B,A b represents the amplitude of the bth power frequency harmonic noise, f 0 Indicates the fundamental frequency of the power frequency harmonics, Indicates the phase of the bth power frequency harmonic noise.

[0026] Optionally, constructing an output vector and a random noise vector according to the ground nuclear magnetic resonance data specifically includes:

[0027] Let n = k, k + 1, ..., k + β-1 construct the output variable and random noise vector corresponding to y(n); the output variable is expressed as: y(k) = [y(k), y(k+1), ..., y(k+β-1)] T ; Wherein, y(k) represents the output vector, k is a positive integer greater than or equal to 1, B+1<β≤Nk, and β is a positive integer, and T represents transpose; the random noise vector is expressed as: w(k)=[w(k),w(k+1),…,w(k+β-1)] T ; Wherein, w(k) represents the random noise vector.

[0028] Optionally, the state vector is expressed as: Wherein, x(k) represents the state vector.

[0029] Optionally, the state space equation is expressed as:

[0030] Wherein, A represents the output matrix, and Φ represents the system matrix.

[0031] Optionally, the relaxation time of the SNMR signal is expressed as: The Larmor frequency of the SNMR signal is expressed as: f L =∠λ / 2π;

[0032] In the formula, Re represents the real part, ln represents the logarithm with the constant e as the base, and λ represents an eigenvalue with the smallest absolute value of the real part in the solution of the system matrix.

[0033] The present invention also provides the following solution:

[0034] A ground nuclear magnetic resonance signal parameter extraction system, the system comprising:

[0035] A ground nuclear magnetic resonance data acquisition module is used to acquire ground nuclear magnetic resonance data; the ground nuclear magnetic resonance data includes a plurality of sampling points; each of the sampling points includes SNMR signals, power frequency harmonic noise and random noise;

[0036] A nuclear magnetic signal model building module, used to build a nuclear magnetic signal model according to the ground nuclear magnetic resonance data; the nuclear magnetic signal model includes an SNMR signal sub-model, a power frequency harmonic noise sub-model and a random noise sub-model;

[0037] An output vector and random noise vector construction module, used to construct an output vector and a random noise vector according to the ground nuclear magnetic resonance data;

[0038] A state vector construction module, used to construct a state vector according to the SNMR signal sub-model and the power frequency harmonic noise sub-model;

[0039] A state space equation construction module, used to construct a state space equation according to the output vector, the state vector and the random noise vector;

[0040] A state space equation solving module, used for solving the state space equation to obtain a solution of a system matrix; the solution of the system matrix includes a plurality of eigenvalues;

[0041] The relaxation time and Larmor frequency determination module is used to determine the relaxation time and Larmor frequency of the SNMR signal according to an eigenvalue with the smallest absolute value of the real part in the solution of the system matrix.

[0042] Optionally, the nuclear magnetic signal model building module specifically includes:

[0043] An SNMR signal sub-model construction unit, used to construct an SNMR signal sub-model according to the SNMR signal;

[0044] A power frequency harmonic noise sub-model construction unit, used to construct a power frequency harmonic noise sub-model according to the power frequency harmonic noise;

[0045] The random noise sub-model construction unit is used to construct a random noise sub-model according to the random noise.

[0046] Optionally, the nuclear magnetic signal model is expressed as:

[0047]

[0048] Wherein, y(n) represents the ground NMR data, n=1, 2, ..., N, N represents the total number of the sampling points, and n represents the sequence number of the sampling point. represents the SNMR signal submodel, represents the power frequency harmonic noise sub-model, w(n) represents the random noise sub-model, E 0 represents the initial amplitude of the SNMR signal, represents the relaxation time of the SNMR signal, j represents the imaginary number sign, and f L represents the Larmor frequency of the SNMR signal, A represents the initial phase of the SNMR signal, B represents the total number of power frequency harmonic components, b represents the serial number of the power frequency harmonic component, b=1,2,...,B,A b represents the amplitude of the bth power frequency harmonic noise, f 0 Indicates the fundamental frequency of the power frequency harmonics, Indicates the phase of the bth power frequency harmonic noise.

[0049] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0050] The method and system for extracting parameters of ground nuclear magnetic resonance signals disclosed in the present invention utilize a nuclear magnetic resonance signal model to construct a state-space equation, and by solving the state-space equation, extract two important parameters of the SNMR signal: relaxation time and Larmor frequency. Since the amplitude of the SNMR signal is time-varying, and the state-space equation is a dynamic time-domain model that can process time-varying signals, applying the state-space equation to SNMR signal parameter extraction can improve the accuracy of parameter extraction, and the amount of calculation is much less than that of the maximum likelihood method. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. 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 creative labor.

[0052] Figure 1 It is a flow chart of an embodiment of a method for extracting ground nuclear magnetic resonance signal parameters according to the present invention;

[0053] Figure 2 It is a structural diagram of an embodiment of a ground nuclear magnetic resonance signal parameter extraction system of the present invention. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] The purpose of the present invention is to provide a method and system for extracting parameters of ground nuclear magnetic resonance signals, so as to improve the accuracy of parameter extraction and reduce the amount of calculation.

[0056] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0057] Figure 1 FIG. 1 is a flow chart of an embodiment of a method for extracting ground nuclear magnetic resonance signal parameters according to the present invention. Figure 1 , the ground nuclear magnetic resonance signal parameter extraction method includes:

[0058] Step 101: Acquire ground nuclear magnetic resonance data; the ground nuclear magnetic resonance data includes multiple sampling points; each sampling point includes SNMR signal, power frequency harmonic noise and random noise.

[0059] Before step 101, the following steps are also included:

[0060] A nuclear magnetic resonance water detector is used to obtain measured ground nuclear magnetic resonance observation data (ground nuclear magnetic resonance data). The measured ground nuclear magnetic resonance observation data obtained by the nuclear magnetic resonance water detector mainly includes SNMR signals, power frequency harmonic noise and random noise.

[0061] Step 102: construct a nuclear magnetic resonance signal model according to the ground nuclear magnetic resonance data; the nuclear magnetic resonance signal model includes a SNMR signal sub-model, a power frequency harmonic noise sub-model and a random noise sub-model.

[0062] The step 102 specifically includes:

[0063] Construct a SNMR signal sub-model based on the SNMR signal.

[0064] A power frequency harmonic noise sub-model is constructed based on the power frequency harmonic noise.

[0065] Construct a random noise sub-model based on random noise.

[0066] In step 102, the nuclear magnetic signal model is expressed as:

[0067]

[0068] Where y(n) represents ground NMR data, n = 1, 2, ..., N, N represents the total number of sampling points, and n represents the sequence number of the sampling point. represents the SNMR signal submodel, represents the power frequency harmonic noise sub-model, w(n) represents the random noise sub-model, E 0 represents the initial amplitude of the SNMR signal, e represents the e-index, represents the relaxation time of the SNMR signal, j represents the imaginary number sign, π represents the circumference, and f L represents the Larmor frequency of the SNMR signal, represents the initial phase of the SNMR signal, ∑ represents the sum, B represents the total number of power frequency harmonic components, b represents the serial number of the power frequency harmonic component, b=1,2,...,B,A b represents the amplitude of the bth power frequency harmonic noise, f 0 Indicates the fundamental frequency of the power frequency harmonics, Indicates the phase of the bth power frequency harmonic noise.

[0069] In step 102, it is assumed that the finite sequence (measured ground nuclear magnetic resonance observation data) collected by the nuclear magnetic resonance groundwater detection instrument is y(n), n=1, 2, ..., N, and N is the number of sampling points. Its mathematical model can be expressed as:

[0070]

[0071] in, is the average decay time, also known as relaxation time, which reflects the average porosity of the underground aquifer. L is the Larmor frequency of the geomagnetic field at the measured location, and w(n) is random noise.

[0072] Then build a mixed signal model:

[0073] In the model shown in formula (1), the SNMR signal is a single component with a relaxation time of The frequency is f L The power frequency harmonics are a group of multi-components with infinite relaxation time and frequency bf 0 The SNMR signal and the power frequency harmonics together can form a mixed sinusoidal signal model:

[0074]

[0075] Obviously, the mixed signal s(n) and the random noise w(n) are independent of each other, and the cross-correlation strategy can be used to remove the interference of w(n) on the mixed signal s(n). If the relaxation time and frequency of each component in the mixed signal s(n) can be calculated, the minimum relaxation time is the relaxation time of the SNMR signal. The frequency corresponding to this relaxation time is the Larmor frequency f of the SNMR signal. L .

[0076] Step 103: construct an output vector and a random noise vector according to the ground nuclear magnetic resonance data.

[0077] The step 103 specifically includes:

[0078] Let n = k, k + 1, ..., k + β-1 construct the output variable and random noise vector corresponding to y(n); the output variable is expressed as: y(k) = [y(k), y(k+1), ..., y(k+β-1)] T ; Where y(k) represents the output vector, k is a positive integer greater than or equal to 1, B+1<β≤Nk, and β is a positive integer, and T represents transpose; the random noise vector is expressed as: w(k)=[w(k),w(k+1),…,w(k+β-1)] T ; where w(k) represents a random noise vector.

[0079] Step 104: construct a state vector according to the SNMR signal sub-model and the power frequency harmonic noise sub-model.

[0080] In step 104, the state vector is expressed as: Wherein, x(k) represents the state vector.

[0081] Step 105: Construct a state space equation according to the output vector, the state vector and the random noise vector.

[0082] In step 105, the state space equation is expressed as:

[0083] Where A represents the output matrix and Φ represents the system matrix.

[0084] The state space equation is established in step 103 to step 105, specifically:

[0085] Define two β×1-dimensional vectors, the output vector y(k) and the random noise vector w(k):

[0086] y(k)=[y(k),y(k+1),…,y(k+β-1)] T (3)

[0087] w(k)=[w(k),w(k+1),…,w(k+β-1)] T (4)

[0088] The y(k) in the output vector corresponds to y(n) in model (1), and the output vector y(k) is constructed by taking n=k, k+1,…, k+β-1.

[0089] Since the number of power frequency harmonic components B is unknown in actual acquisition, through experimental analysis, for the SNMR data collected by the actual water finder, the typical value of β is β>1200. From the data analysis of the water finder, due to the limitation of the instrument hardware, the SNMR signal parameters contained in the first 500 sampling data are not accurate, so the typical value of k is k>500.

[0090] Define the state vector x(k) as:

[0091]

[0092] The state vector is established by the expression on the right side of the equal sign of the model shown in formula (2), or it can be said that it is established by removing the random noise part from the expression on the right side of the equal sign of the model shown in formula (1).

[0093] Then the state space equation composed of y(k), w(k) and x(k) is as follows:

[0094]

[0095] Among them, the output matrix A and the system matrix Φ are:

[0096]

[0097]

[0098] Step 106: Solve the state space equation to obtain a solution of the system matrix; the solution of the system matrix includes multiple eigenvalues.

[0099] In step 106, the SNMR signal relaxation time can be obtained by solving the value of the system matrix Φ for the state space equation shown in the above formula (6): Larmor frequency f L And the estimated value of the fundamental frequency of the power frequency harmonic f 0 , specifically:

[0100] The relaxation time and Larmor frequency are solved using the state-space equation combined with cross-correlation denoising:

[0101] The output vector y(k) defines two i×m dimensional Hankel matrices, named the first output matrix and the second output matrix respectively:

[0102]

[0103]

[0104] Wherein, 2i+m<β+1. The first output matrix shown in formula (9) and the second output matrix shown in formula (10) are constructed by the elements in the output vector y(k).

[0105] Similarly, the first random noise matrix and the second random noise matrix of i×m dimensions are defined as:

[0106]

[0107]

[0108] The first random noise matrix shown in formula (11) and the second random noise matrix shown in formula (12) are constructed from the elements in the random noise vector w(k).

[0109] The first state matrix X 1 , the second state matrix X 2 They are defined as:

[0110] X 1 =[x(k) x(k+1) … x(k+m-1)] (13)

[0111] X 2 =[x(k+i) x(k+i+1) … x(k+i+m-1)] (14)

[0112] The first state matrix X 1 and the second state matrix X 2 Constructed by formula (5).

[0113] The observable matrix Γ is defined as follows:

[0114]

[0115] The second output matrix Y defined above 2 , the second state matrix X 2 , the observable matrix Γ and the second measurement noise matrix (second random noise matrix) W 2 The following equation can be obtained:

[0116] Y 2 =ΓX 2 +W 2 (16)

[0117] Formula (16) is obtained by substituting the elements in the previous definitions into the formula, and finding that the left and right sides are equal. Formula (16) is used to solve formula (18).

[0118] The existing cross-correlation operation method is used to perform cross-correlation operation on formula (16) and the first output matrix. Since the second random noise matrix and the first output matrix are uncorrelated, the result after the cross-correlation operation is:

[0119]

[0120] Performing singular value decomposition on the left side of the equal sign in formula (17) can obtain the left singular matrix, the right singular matrix and the singular value matrix, where the first B+1 columns of the left singular matrix are the left signal singular matrix, which is represented by U 1 Indicates that the matrix composed of the larger B+1 singular values ​​in the singular value matrix is ​​the signal singular value matrix, and Σ 1 From formula (18), we can get Γ and U 1 and Σ 1 The relationship is:

[0121]

[0122] Formula (18) is derived from the properties of singular value decomposition.

[0123] According to the properties of the state space equation, solving the following linear equation can obtain the system matrix Φ after the cross-correlation operation: T :

[0124] Γ(1:β(i-1),1:B+1)Φ T =Γ(β+1:iβ,1:B+1) (19)

[0125] Formula (19) is obtained from the properties of the state space equation. The left side of the equal sign of formula (18) is a matrix, and the elements in it are written as Γ(1:β(i-1),1:B+1) and Γ(β+1:iβ,1:B+1).

[0126] P T Performing eigenvalue decomposition can obtain B+1 eigenvalues ​​λ 0 ,λ 1 ,…λ B , due to Φ T and Φ are similar transformation relations, so the eigenvalues ​​of the two are the same, that is, the solution of the system matrix Φ is: Φ=diag[λ 0 λ 1 …λ B ].

[0127] The above process of solving the state-space equation to obtain the solution of the system matrix is ​​a commonly used process (method) for solving the state-space equation and belongs to the prior art.

[0128] Step 107: Determine the relaxation time of the SNMR signal and the Larmor frequency of the SNMR signal according to an eigenvalue with the smallest absolute value of the real part in the solution of the system matrix.

[0129] In step 107, the relaxation time of the SNMR signal is expressed as: The Larmor frequency of the SNMR signal is expressed as: f L =∠λ / 2π;

[0130] Where Re represents the real part, ln represents the logarithm with the constant e as the base, and λ represents an eigenvalue with the smallest absolute value of the real part in the solution of the system matrix.

[0131] In step 107, λ 0 ,λ 1 ,…λ B The eigenvalue with the smallest absolute value of the real part is denoted by λ, which corresponds to The relaxation time of the SNMR signal is and the Larmor frequency f L It can be obtained by the following formula:

[0132]

[0133] and f L It can be solved by formula (20). The method of calculating the fundamental frequency of the power frequency harmonic is similar. If the e index is known, the logarithm is calculated and then divided by 2π, then f can be obtained. 0 .

[0134] The present invention proposes a method for extracting ground nuclear magnetic resonance signal parameters based on state-space equations. On the basis of the SNMR data model collected by the existing nuclear magnetic resonance water detector, the present invention establishes an SNMR state-space equation, and extracts two important parameters of the SNMR signal: relaxation time and Larmor frequency by solving the state-space equation.

[0135] The present invention combines the SNMR signal and the power frequency harmonic noise in the ground nuclear magnetic resonance observation data collected by the nuclear magnetic resonance water detector into a mixed signal, and uses the state space equation combined with the cross-correlation denoising to solve the characteristic parameters of the SNMR signal. The advantages are:

[0136] 1. Good denoising ability. The SNMR signal and the power frequency harmonic noise form a mixed signal, that is, a mixed signal model is established. The mixed signal and the random noise are independent of each other, which is conducive to removing the influence of random noise by the cross-correlation method (cross-correlation operation). In addition, the power frequency harmonics are regarded as part of the mixed signal and do not need to be filtered out in advance. Even time-varying power frequency harmonics or power frequency harmonics of multiple fundamental frequencies can be processed by the present invention.

[0137] 2. High precision. Since the amplitude of SNMR signal varies with time, and the state space equation is a dynamic time domain model that can process time-varying signals, applying the state space equation to SNMR signal parameter extraction can improve the accuracy of parameter extraction. The extracted SNMR signal parameters is the average decay time, also known as relaxation time, which reflects the average porosity of the underground aquifer. L is the Larmor frequency of the geomagnetic field at the measurement location.

[0138] The present invention adopts state-space equation combined with cross-correlation denoising to extract SNMR signal parameters. The extraction accuracy of relaxation time and Larmor frequency is higher than the method of denoising first and then fitting, and is also higher than the correlation method or high-order statistics method such as ESPRIT. The amount of calculation is much smaller than the maximum likelihood method.

[0139] Figure 2 FIG. 1 is a structural diagram of an embodiment of a ground NMR signal parameter extraction system of the present invention. Figure 2 , the ground nuclear magnetic resonance signal parameter extraction system comprises:

[0140] The ground nuclear magnetic resonance data acquisition module 201 is used to acquire ground nuclear magnetic resonance data; the ground nuclear magnetic resonance data includes multiple sampling points; each of the sampling points includes SNMR signals, power frequency harmonic noise and random noise.

[0141] The nuclear magnetic signal model building module 202 is used to build a nuclear magnetic signal model according to ground nuclear magnetic resonance data; the nuclear magnetic signal model includes a SNMR signal sub-model, a power frequency harmonic noise sub-model and a random noise sub-model.

[0142] The nuclear magnetic signal model building module 202 specifically includes:

[0143] The SNMR signal sub-model construction unit is used to construct the SNMR signal sub-model according to the SNMR signal.

[0144] The power frequency harmonic noise sub-model construction unit is used to construct a power frequency harmonic noise sub-model according to the power frequency harmonic noise.

[0145] The random noise sub-model building unit is used to build a random noise sub-model according to the random noise.

[0146] Among them, the nuclear magnetic signal model is expressed as:

[0147]

[0148] Where y(n) represents ground NMR data, n = 1, 2, ..., N, N represents the total number of sampling points, and n represents the sequence number of the sampling point. represents the SNMR signal submodel, represents the power frequency harmonic noise sub-model, w(n) represents the random noise sub-model, E 0 represents the initial amplitude of the SNMR signal, e represents the e-index, represents the relaxation time of the SNMR signal, j represents the imaginary number sign, π represents the circumference, and f L represents the Larmor frequency of the SNMR signal, represents the initial phase of the SNMR signal, ∑ represents the sum, B represents the total number of power frequency harmonic components, b represents the serial number of the power frequency harmonic component, b=1,2,...,B,A b represents the amplitude of the bth power frequency harmonic noise, f 0 Indicates the fundamental frequency of the power frequency harmonics, Indicates the phase of the bth power frequency harmonic noise.

[0149] The output vector and random noise vector construction module 203 is used to construct an output vector and a random noise vector according to the ground nuclear magnetic resonance data.

[0150] The output vector and random noise vector construction module 203 specifically includes:

[0151] The output vector and random noise vector construction unit is used to construct the output variable and random noise vector corresponding to y(n) by setting n=k, k+1,…,k+β-1; the output variable is expressed as: y(k)=[y(k),y(k+1),…,y(k+β-1)] T ; Where y(k) represents the output vector, k is a positive integer greater than or equal to 1, B+1<β≤Nk, and β is a positive integer, and T represents transpose; the random noise vector is expressed as: w(k)=[w(k),w(k+1),…,w(k+β-1)] T ; where w(k) represents a random noise vector.

[0152] The state vector construction module 204 is used to construct a state vector according to the SNMR signal sub-model and the power frequency harmonic noise sub-model.

[0153] Among them, the state vector is expressed as: Among them, x(k) represents the state vector.

[0154] The state space equation construction module 205 is used to construct the state space equation according to the output vector, the state vector and the random noise vector.

[0155] The state space equation is expressed as:

[0156] Where A represents the output matrix and Φ represents the system matrix.

[0157] The state space equation solving module 206 is used to solve the state space equation to obtain a solution of the system matrix; the solution of the system matrix includes multiple eigenvalues.

[0158] The relaxation time and Larmor frequency determination module 207 is used to determine the relaxation time and Larmor frequency of the SNMR signal according to an eigenvalue with the smallest absolute value of the real part in the solution of the system matrix.

[0159] The relaxation time of the SNMR signal is expressed as: The Larmor frequency of the SNMR signal is expressed as: f L =∠λ / 2π;

[0160] Where Re represents the real part, ln represents the logarithm with the constant e as the base, and λ represents an eigenvalue with the smallest absolute value of the real part in the solution of the system matrix.

[0161] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0162] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for extracting ground nuclear magnetic resonance signal parameters, characterized in that: The method comprises: Acquire ground nuclear magnetic resonance data; the ground nuclear magnetic resonance data includes a plurality of sampling points; each of the sampling points includes SNMR signals, power frequency harmonic noise and random noise; A nuclear magnetic resonance signal model is constructed according to the ground nuclear magnetic resonance data; the nuclear magnetic resonance signal model includes a SNMR signal sub-model, a power frequency harmonic noise sub-model and a random noise sub-model; the nuclear magnetic resonance signal model is expressed as: Wherein, y(n) represents the ground NMR data, n=1, 2, ..., N, N represents the total number of the sampling points, and n represents the sequence number of the sampling point. represents the SNMR signal submodel, represents the power frequency harmonic noise sub-model, w(n) represents the random noise sub-model, E0 represents the initial amplitude of the SNMR signal, represents the relaxation time of the SNMR signal, j represents the imaginary number sign, and f L represents the Larmor frequency of the SNMR signal, A represents the initial phase of the SNMR signal, B represents the total number of power frequency harmonic components, b represents the serial number of the power frequency harmonic component, b=1,2,...,B,A b represents the amplitude of the bth power frequency harmonic noise, f0 represents the power frequency harmonic fundamental frequency, Indicates the phase of the bth power frequency harmonic noise; constructing an output vector and a random noise vector according to the ground nuclear magnetic resonance data; Constructing a state vector according to the SNMR signal sub-model and the power frequency harmonic noise sub-model; constructing a state space equation according to the output vector, the state vector and the random noise vector; Solving the state space equation to obtain a solution of a system matrix; the solution of the system matrix includes a plurality of eigenvalues; The relaxation time of the SNMR signal and the Larmor frequency of the SNMR signal are determined according to an eigenvalue with the smallest absolute value of the real part in the solution of the system matrix; the relaxation time of the SNMR signal is expressed as: The Larmor frequency of the SNMR signal is expressed as: f L =∠λ / 2π; wherein Re represents the real part, ln represents the logarithm with the constant e as the base, and λ represents an eigenvalue with the smallest absolute value of the real part in the solution of the system matrix.

2. The method for extracting ground nuclear magnetic resonance signal parameters according to claim 1, characterized in that: The step of constructing a nuclear magnetic resonance signal model according to the ground nuclear magnetic resonance data specifically includes: Constructing a SNMR signal sub-model according to the SNMR signal; Constructing a power frequency harmonic noise sub-model according to the power frequency harmonic noise; A random noise sub-model is constructed according to the random noise.

3. The method for extracting ground nuclear magnetic resonance signal parameters according to claim 1, characterized in that: The step of constructing an output vector and a random noise vector according to the ground nuclear magnetic resonance data specifically includes: Let n = k, k + 1, ..., k + β-1 construct the output variable and random noise vector corresponding to y(n); the output variable is expressed as: y(k) = [y(k), y(k+1), L, y(k+β-1)] T ; Wherein, y(k) represents the output vector, k is a positive integer greater than or equal to 1, B+1<β≤Nk, and β is a positive integer, and T represents transpose; the random noise vector is expressed as: w(k)=[w(k),w(k+1),L,w(k+β-1)] T ; Wherein, w(k) represents the random noise vector.

4. The method for extracting ground nuclear magnetic resonance signal parameters according to claim 3, characterized in that: The state vector is expressed as: Wherein, x(k) represents the state vector.

5. The method for extracting ground nuclear magnetic resonance signal parameters according to claim 4, characterized in that: The state space equation is expressed as: Wherein, A represents the output matrix, and Φ represents the system matrix.

6. A ground nuclear magnetic resonance signal parameter extraction system, characterized in that: The system comprises: A ground nuclear magnetic resonance data acquisition module is used to acquire ground nuclear magnetic resonance data; the ground nuclear magnetic resonance data includes a plurality of sampling points; each of the sampling points includes SNMR signals, power frequency harmonic noise and random noise; A nuclear magnetic signal model building module is used to build a nuclear magnetic signal model according to the ground nuclear magnetic resonance data; the nuclear magnetic signal model includes an SNMR signal sub-model, a power frequency harmonic noise sub-model and a random noise sub-model; the nuclear magnetic signal model is expressed as: Wherein, y(n) represents the ground NMR data, n=1, 2, ..., N, N represents the total number of the sampling points, and n represents the sequence number of the sampling point. represents the SNMR signal submodel, represents the power frequency harmonic noise sub-model, w(n) represents the random noise sub-model, E0 represents the initial amplitude of the SNMR signal, represents the relaxation time of the SNMR signal, j represents the imaginary number sign, and f L represents the Larmor frequency of the SNMR signal, A represents the initial phase of the SNMR signal, B represents the total number of power frequency harmonic components, b represents the serial number of the power frequency harmonic component, b=1,2,...,B,A b represents the amplitude of the bth power frequency harmonic noise, f0 represents the power frequency harmonic fundamental frequency, Indicates the phase of the bth power frequency harmonic noise; An output vector and random noise vector construction module, used to construct an output vector and a random noise vector according to the ground nuclear magnetic resonance data; A state vector construction module, used to construct a state vector according to the SNMR signal sub-model and the power frequency harmonic noise sub-model; A state space equation construction module, used to construct a state space equation according to the output vector, the state vector and the random noise vector; A state space equation solving module, used for solving the state space equation to obtain a solution of a system matrix; the solution of the system matrix includes a plurality of eigenvalues; The relaxation time and Larmor frequency determination module is used to determine the relaxation time and Larmor frequency of the SNMR signal according to an eigenvalue with the smallest absolute value of the real part in the solution of the system matrix; the relaxation time of the SNMR signal is expressed as: The Larmor frequency of the SNMR signal is expressed as: f L =∠λ / 2π; wherein Re represents the real part, ln represents the logarithm with the constant e as the base, and λ represents an eigenvalue with the smallest absolute value of the real part in the solution of the system matrix.

7. The ground nuclear magnetic resonance signal parameter extraction system according to claim 6, characterized in that: The nuclear magnetic signal model building module specifically includes: An SNMR signal sub-model construction unit, used to construct an SNMR signal sub-model according to the SNMR signal; A power frequency harmonic noise sub-model construction unit, used to construct a power frequency harmonic noise sub-model according to the power frequency harmonic noise; The random noise sub-model construction unit is used to construct a random noise sub-model according to the random noise.

Images