Magnetic anomaly target detection method combining singular spectrum analysis and orthogonal basis method

Abstract

The present application provides a kind of singular spectrum analysis and orthogonal base method combined magnetic anomaly target detection method, device, equipment and storage medium, the method comprises: obtaining the magnetic measurement data to be detected, the magnetic measurement data to be detected is preprocessed;Based on singular spectrum analysis method, the preprocessed magnetic measurement data to be detected is denoised, and denoised signal data is obtained;Based on orthogonal base detection method, the denoised signal data is abnormally signal detected.The present application uses singular spectrum analysis method to first transform one-dimensional measurement data into two-dimensional data to construct trajectory matrix, then singular value decomposition is carried out to data, by selecting appropriate data component, reconstruct high signal-to-noise ratio signal, then using orthogonal base detection algorithm judges whether the target abnormal signal is contained in the data to be measured;So as to greatly improve the magnetic anomaly signal detection ability under low signal-to-noise ratio.

Description

Technical Field

[0001] This invention relates to the field of airborne magnetic detection technology, and in particular to a method, apparatus, equipment, and storage medium for detecting magnetic anomalies by combining singular spectrum analysis and orthogonal basis methods. Background Technology

[0002] Magnetic anomaly signal detection methods extract weak target magnetic signals from complex background magnetic noise, thereby detecting ferromagnetic targets. This method has been widely used in resource exploration, shipwreck salvage, and unexploded ordnance detection. In practical applications, magnetic anomaly signal detection faces two main challenges: First, the background magnetic noise is complex. In stationary platform detection, magnetic sensors are mainly affected by surrounding human-induced magnetic interference and diurnal variations; in aerial platform detection, they are also affected by platform magnetic interference and geological magnetic interference. These magnetic interferences are all colored noise, which not only disrupts the target magnetic anomaly signal in waveform but also has a relatively large amplitude. Second, according to physics, the amplitude of the magnetic signal generated by the target decreases cubically with increasing distance between the target and the measuring sensor. Therefore, as the distance increases, the target signal rapidly decreases and is submerged in noise. Since the orthogonal basis function method, based on the magnetic dipole magnetic field formula, can derive the target magnetic signal as the product of three orthogonal basis functions and their correlation coefficients, and the sum of squared coefficients obtained by matching the orthogonal basis functions with measured magnetic data constitutes the detector, it is often used for target signal detection. However, complex magnetic interference noise reduces its detection capability, so noise suppression algorithms need to be introduced to improve its detection capability. Summary of the Invention

[0003] The present invention aims to provide a method, apparatus, device and storage medium for detecting magnetic anomalies by combining singular spectrum analysis and orthogonal basis method, so as to solve the above-mentioned technical problems and thus effectively improve the ability to detect magnetic anomaly signals in complex magnetic interference noise environment.

[0004] To address the aforementioned technical problems, this invention provides a method for detecting magnetic anomalies that combines singular spectral analysis and orthogonal basis methods, comprising:

[0005] Acquire the magnetic measurement data to be tested, and preprocess the magnetic measurement data to be tested;

[0006] The noise reduction process is performed on the preprocessed magnetic measurement data to be tested based on the singular spectrum analysis method to obtain the noise-reduced signal data.

[0007] Anomaly detection is performed on the denoised signal data based on the orthogonal basis detection method.

[0008] Further, the acquisition of the magnetic measurement data to be tested and the preprocessing of the magnetic measurement data to be tested include:

[0009] Acquire the magnetic measurement data to be tested, and perform high-pass filtering on the magnetic measurement data to be tested using a preset filter.

[0010] Furthermore, the preset filter is a 4th-order high-pass Butterworth filter, and the cutoff frequency of the 4th-order high-pass Butterworth filter is set to 0.025Hz.

[0011] Furthermore, the preprocessed magnetic measurement data to be tested is denoised using the singular spectrum analysis method to obtain denoised signal data, including:

[0012] Based on a pre-selected data window length, the magnetic measurement data to be tested is rearranged in the form of a sliding window to construct a trajectory matrix;

[0013] Singular value decomposition is performed on the trajectory matrix to obtain the singular values ​​and corresponding eigenvectors corresponding to the trajectory matrix.

[0014] The matrix components of the trajectory matrix are constructed based on the decomposed singular values ​​and their corresponding eigenvectors.

[0015] Based on the distribution of the singular values, a preset clustering algorithm is used to classify all matrix components of the trajectory matrix, and the matrix components of the target category are selected according to a preset strategy. The denoised signal data is then reconstructed based on the matrix components of the target category.

[0016] Further, based on the distribution of the singular values, a preset clustering algorithm is used to classify all matrix components of the trajectory matrix, and matrix components of the target category are selected according to a preset strategy. The denoised signal data is then reconstructed based on the matrix components of the target category, specifically including:

[0017] Based on the distribution of the singular values, a preset clustering algorithm is used to divide all matrix components of the trajectory matrix into three categories;

[0018] According to a preset strategy, the matrix components of the target category after removing low-frequency components and high-frequency noise are selected from the matrix components of the three categories, and the noise-reduced signal data is obtained by reconstructing based on the matrix components of the target category.

[0019] Furthermore, the abnormal signal detection of the denoised signal data based on the orthogonal basis detection method includes:

[0020] Construct orthogonal basis detection functions, and use the orthogonal basis functions to match the denoised signal data to obtain matching coefficients;

[0021] The energy function is determined based on the matching coefficients;

[0022] The energy threshold is determined based on the energy corresponding to the preset historical time period prior to the current moment.

[0023] The energy of the denoised signal data is obtained using the energy function, and the presence of a target anomalous signal in the magnetic measurement data under test is determined based on the comparison between the energy of the denoised signal data and the energy threshold.

[0024] Furthermore, determining whether the magnetic measurement data under test contains a target anomalous signal based on the comparison result between the energy of the noise-reduced signal data and the energy threshold specifically includes:

[0025] If the energy of the noise-reduced signal data is greater than the energy threshold, and the energy of the noise-reduced signal data reaches the peak point, then it is determined that there is a target abnormal signal in the magnetic measurement data to be tested at the current moment.

[0026] If the energy of the noise-reduced signal data is not greater than the energy threshold, or the energy of the noise-reduced signal data has not reached the peak point, then it is determined that there is no target abnormal signal in the magnetic measurement data to be tested at the current moment.

[0027] This invention also provides a magnetic anomaly target detection device combining singular spectrum analysis and orthogonal basis methods, comprising:

[0028] The preprocessing module is used to acquire the magnetic measurement data to be tested and to preprocess the magnetic measurement data to be tested.

[0029] The noise reduction module is used to perform noise reduction processing on the preprocessed magnetic measurement data to be tested based on the singular spectrum analysis method to obtain noise-reduced signal data.

[0030] The detection module is used to detect abnormal signals in the noise-reduced signal data based on the orthogonal basis detection method.

[0031] The present invention also provides a terminal device, including a processor and a memory storing a computer program, wherein the processor executes the computer program to implement the magnetic anomaly target detection method combining singular spectrum analysis and orthogonal basis method as described in any one of the claims.

[0032] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the magnetic anomaly target detection method combining singular spectrum analysis and orthogonal basis method as described in any one of the claims.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] This invention provides a method, apparatus, device, and storage medium for detecting magnetic anomalies by combining singular spectral analysis and orthogonal basis methods. The method includes: acquiring magnetic measurement data to be detected; preprocessing the magnetic measurement data; performing noise reduction processing on the preprocessed magnetic measurement data based on singular spectral analysis to obtain denoised signal data; and detecting anomalies in the denoised signal data based on orthogonal basis detection. This invention utilizes singular spectral analysis to first transform one-dimensional measurement data into two-dimensional data to construct a trajectory matrix, then performs singular value decomposition on the data, and reconstructs a high signal-to-noise ratio signal by selecting appropriate data components. Finally, it uses an orthogonal basis detection algorithm to determine whether the measured data contains a target anomaly signal, thereby significantly improving the detection capability of magnetic anomalies under low signal-to-noise ratio conditions. Attached Figure Description

[0035] Figure 1 This is one of the flowcharts of the magnetic anomaly target detection method combining singular spectrum analysis and orthogonal basis method provided by the present invention;

[0036] Figure 2 This is the second flowchart of the magnetic anomaly target detection method combining singular spectrum analysis and orthogonal basis method provided by the present invention;

[0037] Figure 3 This is a schematic diagram of the magnetic anomaly target detection device that combines singular spectrum analysis and orthogonal basis method provided by the present invention. Detailed Implementation

[0038] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0039] Please see Figure 1 This invention provides a method for detecting magnetic anomalies by combining singular spectral analysis and orthogonal basis methods, which may include the following steps:

[0040] S1. Obtain the magnetic measurement data to be tested, and preprocess the magnetic measurement data to be tested;

[0041] S2. Based on the singular spectrum analysis method, the preprocessed magnetic measurement data to be tested is denoised to obtain denoised signal data.

[0042] S3. Anomaly detection is performed on the denoised signal data based on the orthogonal basis detection method.

[0043] In this embodiment of the invention, step S1 specifically includes:

[0044] Acquire the magnetic measurement data to be tested, and perform high-pass filtering on the magnetic measurement data to be tested using a preset filter.

[0045] In this embodiment of the invention, the preset filter is further defined as a 4th-order high-pass Butterworth filter, and the cutoff frequency of the 4th-order high-pass Butterworth filter is set to 0.025Hz.

[0046] In this embodiment of the invention, step S2 specifically includes:

[0047] Based on a pre-selected data window length, the magnetic measurement data to be tested is rearranged in the form of a sliding window to construct a trajectory matrix;

[0048] Singular value decomposition is performed on the trajectory matrix to obtain the singular values ​​and corresponding eigenvectors corresponding to the trajectory matrix.

[0049] The matrix components of the trajectory matrix are constructed based on the decomposed singular values ​​and their corresponding eigenvectors.

[0050] Based on the distribution of the singular values, a preset clustering algorithm is used to classify all matrix components of the trajectory matrix, and the matrix components of the target category are selected according to a preset strategy. The denoised signal data is then reconstructed based on the matrix components of the target category.

[0051] In this embodiment of the invention, further, based on the distribution of the singular values, a preset clustering algorithm is used to classify all matrix components of the trajectory matrix, and matrix components of the target category are selected according to a preset strategy. The denoised signal data is then reconstructed based on the matrix components of the target category. Specifically, this includes:

[0052] Based on the distribution of the singular values, a preset clustering algorithm is used to divide all matrix components of the trajectory matrix into three categories;

[0053] According to a preset strategy, the matrix components of the target category after removing low-frequency components and high-frequency noise are selected from the matrix components of the three categories, and the noise-reduced signal data is obtained by reconstructing based on the matrix components of the target category.

[0054] In this embodiment of the invention, step S3 specifically includes:

[0055] Construct orthogonal basis detection functions, and use the orthogonal basis functions to match the denoised signal data to obtain matching coefficients;

[0056] The energy function is determined based on the matching coefficients;

[0057] The energy threshold is determined based on the energy corresponding to the preset historical time period prior to the current moment.

[0058] The energy of the denoised signal data is obtained using the energy function, and the presence of a target anomalous signal in the magnetic measurement data under test is determined based on the comparison between the energy of the denoised signal data and the energy threshold.

[0059] In this embodiment of the invention, further, determining whether there is a target anomalous signal in the magnetic measurement data to be tested based on the comparison result between the energy of the noise-reduced signal data and the energy threshold specifically includes:

[0060] If the energy of the noise-reduced signal data is greater than the energy threshold, and the energy of the noise-reduced signal data reaches the peak point, then it is determined that there is a target abnormal signal in the magnetic measurement data to be tested at the current moment.

[0061] If the energy of the noise-reduced signal data is not greater than the energy threshold, or the energy of the noise-reduced signal data has not reached the peak point, then it is determined that there is no target abnormal signal in the magnetic measurement data to be tested at the current moment.

[0062] Based on the above scheme, to facilitate a better understanding of the magnetic anomaly target detection method combining singular spectral analysis and orthogonal basis methods provided in the embodiments of the present invention, the following detailed description is provided:

[0063] It should be noted that, in this embodiment of the invention, a suitable window length is first selected for the data to be measured. Then, the one-dimensional measurement signal is converted into a two-dimensional trajectory matrix by sliding window. Then, the matrix is ​​decomposed into singular values ​​and eigenvectors by singular value analysis. By analyzing the distribution law of singular values, suitable singular values ​​are selected to reconstruct a measurement signal with a high signal-to-noise ratio. Then, the reconstructed signal is detected by orthogonal basis method to determine whether the measurement data contains a target anomalous signal.

[0064] The embodiments of the present invention can be implemented through the following steps:

[0065] 1) Load the magnetic measurement data to be detected and preprocess the data;

[0066] 2) Obtain the denoised data containing the target signal using singular spectrum analysis;

[0067] 3) Use orthogonal basis detection method to determine whether the denoised data contains the target signal.

[0068] The following is a detailed explanation.

[0069] First, load the magnetic measurement data to be detected. For real-time measurement data, a segmented detection method can be adopted, but it should be noted that the length of each segment should be greater than the length of the signal to be detected. Then, use a high-pass filter to filter the measurement data.

[0070] Next, use the singular spectrum analysis method to process the measurement data. The specific steps are as follows:

[0071] (1) Assume that B m is the measurement data. Let B m = [x1, x2,..., x N , where N is the sequence length. First, select an appropriate window length L, and rearrange B m to obtain the trajectory matrix X.

[0072]

[0073] Usually, take L < N / 2. Let K = N - L + 1, then x can be expressed as:

[0074]

[0075] (2) Perform singular value decomposition (SVD) on X. First, calculate the covariance matrix of x:

[0076] S = XX T (3)

[0077] Then, perform eigenvalue decomposition on S, and the eigenvalues λ1, λ2... λ L and the corresponding eigenvectors δ1, δ2,..., δ L can be obtained. Here, the eigenvalues and the corresponding eigenvectors are arranged in descending order of the eigenvalues. Let δ = [δ1, δ2,..., δ L , At this time, the m-th component of the trajectory matrix x can be obtained:

[0078] <,

[0079] Simplify the above formula to get: <,

[0080]

[0081] where represents a subspace composed of eigenvectors. Project x onto this subspace to obtain the m-th component of the trajectory matrix x. L components can be obtained from the above steps.

[0082] (3) Group the above L components into c groups. Let O = [g1, g2, ..., g u The ] indicates that u singular values ​​were selected from the decomposition results for combination and reconstruction. Then the trajectory matrix X of group O... O It can be represented as:

[0083]

[0084] The original trajectory matrix X can then be represented as:

[0085]

[0086] In this embodiment of the invention, a clustering method is used to classify the obtained singular values. Each singular value corresponds one-to-one with its characteristic component, and the distribution of the singular values ​​characterizes the properties of that characteristic component to a certain extent. All singular values ​​are classified according to their magnitude and distribution. Analysis shows that the signals contained in the measurement data can be divided into three categories: low-frequency signals representing trends, high-frequency components representing near-Gaussian noise, and intermediate components interspersed with signals. The singular values ​​corresponding to low-frequency components are larger, while those corresponding to Gaussian noise are smaller. Therefore, a clustering method is used to decompose the obtained singular values ​​into three categories, removing low-frequency and high-frequency components.

[0087] (4) Transform the new trajectory matrix obtained in the previous step into a new sequence of length N. Let Y be an L×K matrix, where y ij Let l be the element in the i-th row and j-th column of the matrix. min =min{L,K},l max =max{L,K}, then Y can be transformed into The one-dimensional time series is as follows:

[0088]

[0089] At this point, a new denoised sequence Y (denoised signal data) can be obtained, in which most of the noise has been removed.

[0090] Next, the orthogonal basis detection method is used to detect the signal in the denoised quasi-target data. The specific steps are as follows:

[0091] (1) The orthogonal basis function detection method is based on the magnetic dipole theory formula. This formula is derived to obtain a detection method containing three orthogonal basis functions. These three basis functions are mutually orthogonal and normalized, thus exhibiting high resistance to Gaussian white noise. The magnetic signal formula derived from the magnetic dipole magnetic field formula can be written as:

[0092]

[0093] in:

[0094]

[0095]

[0096]

[0097]

[0098] in,

[0099]

[0100] Based on the orthogonality of basis functions, the coefficient a j The values ​​of j = 1, 2, 3 can be obtained using the following formula:

[0101]

[0102] In actual detection, the data collected by the magnetometer is discrete data. Therefore, the continuous integral form of the formula should be transformed into a discrete summation form, and its expression is:

[0103]

[0104] Where Δw represents the length of the spatial sampling, and k represents the half-window length, typically taken as 3. During the detection process, a point-by-point detection method can be used, where m represents the current detection point.

[0105] (2) Define the energy function for constructing the detector.

[0106]

[0107] (3) Determine the threshold to determine whether the target is present. In practical applications, the threshold is usually selected using a constant false alarm strategy, which is set by using a fixed multiple of the sum of the energy of the previous segment of the current data.

[0108]

[0109] In the above formula, m represents the current value, ξ represents the interval, which is generally half the signal width; p is half the signal width, and η represents the multiple greater than the noise energy, which determines the false alarm rate and the missed alarm rate.

[0110] When E(m) > T res old(m) and E(m) is the peak point, it can be determined that there is a suspected target signal at this time.

[0111] The technical solution of this invention can be implemented through the following steps:

[0112] 1. Perform high-pass filtering on the measurement data to remove out-of-band low-frequency signals. The filter used here is a 4th-order high-pass Butterworth filter with a cutoff frequency set to 0.025Hz.

[0113] 2. Perform singular spectrum analysis on the measurement data, specifically including:

[0114] 2A. First, select an appropriate data length and construct the trajectory matrix X using a sliding window.

[0115] Let B m To measure the data, let B m =[x1,x2,…,x N ], N is the sequence length, and L is the selected window length.

[0116]

[0117] 2B. Perform singular value decomposition on the above trajectory matrix X to obtain the corresponding singular values ​​and corresponding eigenvectors;

[0118] S = XX T Formula 2

[0119] [λ,δ]=eig(XX T Equation 3 In the above formula, eig represents eigenvalue decomposition.

[0120] 2C. Construct a new trajectory matrix using each singular value after decomposition and its corresponding eigenvector, which can then be used as a component of the trajectory matrix X;

[0121] λ=[λ1,λ2…λ L Equation 4

[0122] δ=[δ1,δ2,…,δ L Equation 5

[0123] Then we can get:

[0124]

[0125] in, Let X represent a subspace composed of eigenvectors. Projecting X onto this subspace will yield the m-th component of the trajectory matrix X. The above steps will yield L components.

[0126] 2D. Utilizing the singular value distribution, a clustering algorithm is used to classify the above L components, select appropriate categories, and construct a new trajectory matrix.

[0127] Reconstruct the data by selecting appropriate categories, let o = [g1, g2, ..., g i] indicates that u singular values ​​were selected from the decomposition for combination and reconstruction. Then the trajectory matrix X of group o... O It can be represented as:

[0128]

[0129] 2E. Reconstruct the denoised data Y using the new trajectory matrix.

[0130] Let Y be an L×K matrix, where y ij Let l be the element in the i-th row and j-th column of the matrix. min =min{L,K},l max =max{L,K}, then Y can be transformed into The one-dimensional time series is obtained.

[0131]

[0132] 3. Use the orthogonal basis detection algorithm to detect and identify the denoised data. Specific steps are as follows:

[0133] 3A. Construct orthogonal basis detection functions;

[0134] By selecting a suitable characteristic time value τ, orthogonal basis functions are constructed. The length of the orthogonal basis functions is usually set to 6·τ·f. s , where f s The sampling rate.

[0135]

[0136] 3B. Using orthogonal basis functions to analyze the denoised measurement data B s Perform matching to obtain the corresponding matching coefficients;

[0137]

[0138] 3C. Obtain the energy function using the matching coefficients;

[0139]

[0140] 3D. Determine the threshold using the energy from a previous time period based on the current data;

[0141]

[0142] 3E. Determine if the target signal is present;

[0143] When E(m) > T res old(m) and E(m) is the peak point, it can be determined that there is a suspected target signal at this time.

[0144] Please see Figure 2The working process and principle of the present invention will be illustrated below with specific embodiments:

[0145] 1. First, measure data B m Perform high-pass filtering. Set highpassfilter to represent the high-pass filter. This high-pass filter is a 4th-order high-pass Butterworth filter with a low cutoff frequency of 0.025Hz.

[0146] B filter =highpassfilter(B m )

[0147] 2. Data B filter The length of B is represented by N. filter = [x1, x2, ..., x N If the sliding window length is chosen to be L = N / 2, the constructed trajectory matrix is ​​shown in the following formula;

[0148]

[0149] 3. Perform singular value decomposition on the trajectory matrix above. First, obtain the covariance matrix S of the trajectory matrix. Then, perform eigenvalue decomposition on the covariance matrix to obtain several eigenvalues ​​(i.e., singular values) and eigenvectors.

[0150] s = XX T

[0151] [λ, δ] = eig(XX) T )

[0152] 4. Using clustering methods, the above singular values ​​are divided into 3 categories. The second category is selected to reconstruct the data and obtain a new trajectory matrix; let o = [g1, g2, ..., g u The ] indicates that u singular values ​​were selected from the decomposition for combination and reconstruction. Then the trajectory matrix X of group O... O It can be represented as:

[0153]

[0154] 5. By adding the diagonals, the denoised data for this segment is obtained using the new trajectory matrix described above.

[0155]

[0156] 6. Construct basis functions using the formula, set an appropriate characteristic time τ value, and set the length of the basis functions to 6·τ·f. s , where f s The sampling rate.

[0157]

[0158] 7. Using the above three orthogonal basis functions to analyze the denoised measurement data B s Perform matching to obtain the corresponding coefficients;

[0159]

[0160] In the formula above, k = 3, and m represents the value at the current moment. By sliding the data point by point, all the coefficients of this segment can be obtained.

[0161] 8. Construct the energy function using the coefficients obtained from the matching;

[0162]

[0163] 9. Determine the threshold using the data from previous moments, where η represents the threshold multiple;

[0164]

[0165] 10. Determine if a signal is present:

[0166]

[0167] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0168] When detecting magnetic anomaly signals, the measurement data typically contains anthropogenic magnetic interference, diurnal magnetic noise, and residual magnetic noise compensated by the magnetic sensor platform. Existing detection methods cannot effectively detect target signals directly. This invention proposes using singular spectrum analysis based on nonlinear non-Gaussian signal decomposition to remove noise and improve the detection signal-to-noise ratio (SNR). Then, an orthogonal basis detection method is used to detect the denoised data, thereby determining whether the measurement data contains the target signal. This invention effectively improves the detection capability of magnetic anomaly signals under low SNR conditions.

[0169] It should be noted that, for the sake of simplicity, the above methods or process embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of the present invention are not limited to the described order of actions, because according to the embodiments of the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions involved are not necessarily essential to the embodiments of the present invention.

[0170] Please see Figure 3 This invention also provides a magnetic anomaly target detection device combining singular spectrum analysis and orthogonal basis methods, comprising:

[0171] Preprocessing module 1 is used to acquire the magnetic measurement data to be tested and to preprocess the magnetic measurement data to be tested;

[0172] Noise reduction module 2 is used to perform noise reduction processing on the preprocessed magnetic measurement data to be tested based on the singular spectrum analysis method to obtain noise-reduced signal data;

[0173] The detection module 3 is used to detect abnormal signals in the noise-reduced signal data based on the orthogonal basis detection method.

[0174] Furthermore, the preprocessing module 1 is specifically used for:

[0175] Acquire the magnetic measurement data to be tested, and perform high-pass filtering on the magnetic measurement data to be tested using a preset filter.

[0176] Furthermore, the preset filter is a 4th-order high-pass Butterworth filter, and the cutoff frequency of the 4th-order high-pass Butterworth filter is set to 0.025Hz.

[0177] Furthermore, the noise reduction module 2 is specifically used for:

[0178] Based on a pre-selected data window length, the magnetic measurement data to be tested is rearranged in the form of a sliding window to construct a trajectory matrix;

[0179] Singular value decomposition is performed on the trajectory matrix to obtain the singular values ​​and corresponding eigenvectors corresponding to the trajectory matrix.

[0180] The matrix components of the trajectory matrix are constructed based on the decomposed singular values ​​and their corresponding eigenvectors.

[0181] Based on the distribution of the singular values, a preset clustering algorithm is used to classify all matrix components of the trajectory matrix, and the matrix components of the target category are selected according to a preset strategy. The denoised signal data is then reconstructed based on the matrix components of the target category.

[0182] Furthermore, the noise reduction module 2 is specifically used for:

[0183] Based on the distribution of the singular values, a preset clustering algorithm is used to divide all matrix components of the trajectory matrix into three categories;

[0184] According to a preset strategy, the matrix components of the target category after removing low-frequency components and high-frequency noise are selected from the matrix components of the three categories, and the noise-reduced signal data is obtained by reconstructing based on the matrix components of the target category.

[0185] Furthermore, the detection module 3 is specifically used for:

[0186] Construct orthogonal basis detection functions, and use the orthogonal basis functions to match the denoised signal data to obtain matching coefficients;

[0187] The energy function is determined based on the matching coefficients;

[0188] The energy threshold is determined based on the energy corresponding to the preset historical time period prior to the current moment.

[0189] The energy of the denoised signal data is obtained using the energy function, and the presence of a target anomalous signal in the magnetic measurement data under test is determined based on the comparison between the energy of the denoised signal data and the energy threshold.

[0190] Furthermore, the detection module 3 is specifically used for:

[0191] If the energy of the noise-reduced signal data is greater than the energy threshold, and the energy of the noise-reduced signal data reaches the peak point, then it is determined that there is a target abnormal signal in the magnetic measurement data to be tested at the current moment.

[0192] If the energy of the noise-reduced signal data is not greater than the energy threshold, or the energy of the noise-reduced signal data has not reached the peak point, then it is determined that there is no target abnormal signal in the magnetic measurement data to be tested at the current moment.

[0193] It is understood that the above-described device embodiments correspond to the method embodiments of the present invention. The magnetic anomaly target detection device combining singular spectrum analysis and orthogonal basis method provided by the embodiments of the present invention can realize the magnetic anomaly target detection method combining singular spectrum analysis and orthogonal basis method provided by any one of the method embodiments of the present invention.

[0194] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the magnetic anomaly target detection method combining singular spectrum analysis and orthogonal basis method as described in any one of the claims.

[0195] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0196] Those skilled in the art will clearly understand that, for convenience and brevity, the specific working process of the device described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0197] The terminal device can be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0198] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0199] The memory can be used to store the computer program. The processor implements various functions of the terminal device by running or executing the computer program stored in the memory and calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function, etc.; the data storage area may store data created based on the use of the mobile phone, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, RAM, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0200] The storage medium is a computer-readable storage medium, and the computer program is stored in the computer-readable storage medium. When executed by a processor, the computer program can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0201] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for detecting magnetic anomalies combining singular spectral analysis and orthogonal basis methods, characterized in that, include: Acquire the magnetic measurement data to be tested, and preprocess the magnetic measurement data to be tested; The noise reduction process for preprocessed magnetic measurement data under test based on singular spectral analysis is as follows: The process includes: rearranging the magnetic measurement data under test using a sliding window based on a pre-selected data window length to construct a trajectory matrix; performing singular value decomposition on the trajectory matrix to obtain singular values ​​and corresponding eigenvectors; constructing matrix components of the trajectory matrix based on the decomposed singular values ​​and corresponding eigenvectors; classifying all matrix components of the trajectory matrix using a preset clustering algorithm based on the distribution of the singular values, selecting matrix components of the target category according to a preset strategy, and reconstructing the noise reduction signal data based on the matrix components of the target category. Anomaly detection of the denoised signal data based on the orthogonal basis detection method includes: constructing an orthogonal basis detection function and using the orthogonal basis function to match the denoised signal data to obtain matching coefficients; determining an energy function based on the matching coefficients; determining an energy threshold based on the energy corresponding to a preset historical time period before the current time; calculating the energy of the denoised signal data using the energy function, and determining whether there is a target anomalous signal in the magnetic measurement data to be tested based on the comparison result between the energy of the denoised signal data and the energy threshold.

2. The magnetic anomaly target detection method combining singular spectrum analysis and orthogonal basis method according to claim 1, characterized in that, The process of acquiring the magnetic measurement data to be tested and preprocessing the magnetic measurement data to be tested includes: Acquire the magnetic measurement data to be tested, and perform high-pass filtering on the magnetic measurement data to be tested using a preset filter.

3. The magnetic anomaly target detection method combining singular spectral analysis and orthogonal basis method according to claim 2, characterized in that, The preset filter is a 4th-order high-pass Butterworth filter, and the cutoff frequency of the 4th-order high-pass Butterworth filter is set to 0.025Hz.

4. The magnetic anomaly target detection method combining singular spectral analysis and orthogonal basis method according to claim 1, characterized in that, Based on the distribution of the singular values, a preset clustering algorithm is used to classify all matrix components of the trajectory matrix, and matrix components of the target category are selected according to a preset strategy. The denoised signal data is then reconstructed based on the matrix components of the target category, specifically including: Based on the distribution of the singular values, a preset clustering algorithm is used to divide all matrix components of the trajectory matrix into three categories; According to a preset strategy, the matrix components of the target category after removing low-frequency components and high-frequency noise are selected from the matrix components of the three categories, and the noise-reduced signal data is obtained by reconstructing based on the matrix components of the target category.

5. The magnetic anomaly target detection method combining singular spectrum analysis and orthogonal basis method according to claim 1, characterized in that, The determination of whether the magnetic measurement data under test contains a target anomalous signal based on the comparison result between the energy of the noise-reduced signal data and the energy threshold specifically includes: If the energy of the noise-reduced signal data is greater than the energy threshold, and the energy of the noise-reduced signal data reaches the peak point, then it is determined that there is a target abnormal signal in the magnetic measurement data to be tested at the current moment. If the energy of the noise-reduced signal data is not greater than the energy threshold, or the energy of the noise-reduced signal data has not reached the peak point, then it is determined that there is no target abnormal signal in the magnetic measurement data to be tested at the current moment.

6. A magnetic anomaly target detection device combining singular spectrum analysis and orthogonal basis method, characterized in that, include: The preprocessing module is used to acquire the magnetic measurement data to be tested and to preprocess the magnetic measurement data to be tested. A noise reduction module is used to perform noise reduction processing on preprocessed magnetic measurement data to be tested based on singular spectrum analysis to obtain noise-reduced signal data. This includes: rearranging the magnetic measurement data to be tested using a sliding window based on a pre-selected data window length to construct a trajectory matrix; performing singular value decomposition on the trajectory matrix to obtain singular values ​​and corresponding eigenvectors; constructing matrix components of the trajectory matrix based on each decomposed singular value and its corresponding eigenvector; classifying all matrix components of the trajectory matrix using a preset clustering algorithm based on the distribution of the singular values, selecting the matrix components of the target category according to a preset strategy, and reconstructing the noise-reduced signal data based on the matrix components of the target category. The detection module is used to detect abnormal signals in the denoised signal data based on the orthogonal basis detection method, including: constructing an orthogonal basis detection function and using the orthogonal basis function to match the denoised signal data to obtain matching coefficients; determining an energy function based on the matching coefficients; determining an energy threshold based on the energy corresponding to a preset historical time period before the current time; calculating the energy of the denoised signal data using the energy function, and determining whether there is a target abnormal signal in the magnetic measurement data to be tested based on the comparison result between the energy of the denoised signal data and the energy threshold.

7. A terminal device, comprising a processor and a memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the magnetic anomaly target detection method combining singular spectrum analysis and orthogonal basis method as described in any one of claims 1 to 5.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the magnetic anomaly target detection method combining singular spectrum analysis and orthogonal basis method as described in any one of claims 1 to 5.

Images