Single vector magnetometer moving direction magnetic gradient anomaly detection method

Abstract

The application relates to a single-vector magnetometer motion direction magnetic gradient anomaly detection method, which comprises the following steps: acquiring an x-axis motion direction magnetic gradient signal through spatial sampling based on a single-sensor magnetic gradient measurement technology; pre-processing the x-axis motion direction magnetic gradient signal; presetting an orthogonal base function of the x-axis motion direction magnetic gradient; and performing magnetic gradient anomaly detection on the pre-processed signal based on the preset orthogonal base function. The application is superior to a traditional vertical motion direction array double-sensor column magnetic gradient anomaly OBF detection method under low noise ratio and platform shaking conditions, has the advantages of strong real-time performance, low cost and small size, and shows good application prospects in scene detection, explosive disposal, intrusion monitoring and the like.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of magnetic gradient anomaly detection, in particular to a single vector magnetometer moving direction magnetic gradient anomaly detection method. BACKGROUND

[0002] The magnetic anomaly detection technology captures the disturbance change of the environmental magnetic field by high-precision magnetic sensors and detection algorithms to detect and locate it, and has a wide range of applications in many fields such as military, resource exploration, environmental monitoring, etc. However, since the magnetic anomaly signal attenuation is proportional to the cube of the detection distance, the signal-to-noise ratio of the target magnetic anomaly signal decreases rapidly when the target distance is slightly far. At the same time, the frequency of the target magnetic anomaly signal is low, and it is disturbed by the internal thermal noise of the receiver, the background magnetic field and the platform motion and other complex factors, making it very difficult to detect weak magnetic anomaly signals.

[0003] According to the adopted detection quantities and algorithm principles, the current magnetic anomaly detection techniques can be roughly divided into three categories: the first category is the detection method based on the different statistical characteristics of signals and noises, including the minimum entropy detector and the high-order zero-crossing detector. This kind of method has the advantage of smaller calculation amount compared with the OBF detector, but only uses the target characteristics or noise statistical characteristics to construct the detector, and the detection performance is difficult to achieve the ideal effect in the case of non-Gaussian white noise background noise. The second category is the detection method based on the signal orthogonal basis function (OBF) decomposition, including the scalar magnetic anomaly detection, vector magnetic anomaly detection and magnetic gradient anomaly detection, etc. Ginzburg et al. proposed to decompose the magnetic gradient anomaly signal into a set of orthogonal basis functions, and realize the effective detection of the magnetic anomaly signal through the matching filtering of the signal and the orthogonal basis function template. Hu et al. proposed a model-based vector detection and positioning method of magnetic objects based on the orthogonalization of mother functions, which has high detection and positioning accuracy and stability. The scalar measurement is not affected by the attitude change of the platform, but the high-precision scalar magnetometer is expensive, and the obtained target magnetic field information is limited and the measurement signal is superimposed with various interference magnetic fields; the vector magnetic measurement can obtain the size and direction information of the geomagnetic field, but the slight shaking of the measurement platform will introduce a lot of interference due to the existence of the geomagnetic field; compared with the magnetic scalar detection and magnetic vector detection, the vector magnetic gradient detection measurement result is independent of the local magnetic inclination and magnetic declination, and has strong common-mode rejection ability to the interference introduced by the measurement platform motion and the diurnal variation effect, so with the development of magnetometer technology, the vector magnetic gradient magnetic anomaly detection method is increasingly valued. Jin et al. proposed an OBF detection method based on magnetic tensor contraction, which reduces the influence of noise on target detection and improves the detection accuracy. Ma et al. proposed a vector magnetic gradient anomaly detection method, which improves the stable detection of magnetic anomaly signals under low signal-to-noise ratio and platform shaking conditions. Shen et al. proposed to use the spatial anisotropy characteristics of noise to suppress low-frequency magnetic noise and improve the detection ability of magnetic anomaly signals. This kind of method is simple in calculation and can be used for real-time detection, but it needs multiple scalar or vector magnetometers to form an array, which is high in cost, complicated in correction and does not consider the different influences of array configuration optimization and multi-source error interference on the detection performance. The third category is the weak magnetic signal detection method based on neural network. That is, the magnetic anomaly detection is transformed into the identification and classification problem of magnetic field signal existence or not by using machine learning. Fan et al. proposed a MAD method based on support vector machine, which improves the detection ability of weak magnetic anomaly signals, but needs to manually design the magnetic anomaly signal feature information. Hu et al. proposed a convolutional neural network (CNN) model for magnetic anomaly target signal detection, which lets the network itself extract features for identification and classification, and improves the adaptability of magnetic anomaly detection.Fan et al. first obtain magnetic signal features of different dimensions through pretreatment, then use a Conv1D network to extract feature information of the magnetic anomaly signal, and use the obtained multi-dimensional features for fusion to detect the magnetic anomaly signal. This kind of method needs to collect a large number of training samples, and the quality of the samples is directly related to the performance of the detector. At the same time, the magnetic noise has a great influence on the representation of the target feature of the magnetic anomaly signal, and the detection performance decreases rapidly under low signal-to-noise ratio. SUMMARY

[0004] The purpose of the present application is to propose a single vector magnetometer motion direction magnetic gradient anomaly detection method to solve the problems existing in the prior art.

[0005] To achieve the above purpose, the present application provides the following scheme:

[0006] The single vector magnetometer motion direction magnetic gradient anomaly detection method comprises:

[0007] Based on single sensor magnetic gradient measurement technology, the x-axis motion direction magnetic gradient signal is obtained by spatial sampling;

[0008] The x-axis motion direction magnetic gradient signal is pretreated;

[0009] The orthogonal basis function of the preset x-axis motion direction magnetic gradient is set;

[0010] Based on the preset orthogonal basis function, the pretreated signal is detected for magnetic gradient anomaly.

[0011] Optionally, the x-axis motion direction magnetic gradient signal is:

[0012] G(x)=[B(x+Δx)-B(x)] / Δx

[0013] Wherein, G(x) represents the magnetic gradient signal, B(x+Δx) represents the magnetic induction intensity at x+Δx, B(x) represents the magnetic induction intensity at x, and Δx represents the fixed interval distance.

[0014] Optionally, the pretreatment of the x-axis motion direction magnetic gradient signal comprises:

[0015] The high-frequency noise is suppressed by band-pass filtering.

[0016] Optionally, the preset orthogonal basis function is:

[0017]

[0018] Wherein, The preset orthogonal basis function is represented by ω, and ω represents the ratio of the X-axis direction distance to the CPA distance.

[0019] Optionally, the magnetic gradient anomaly detection comprises:

[0020] The matching degree of the signal after signal preprocessing with the preset orthogonal base function is calculated, and if the matching result exceeds a set threshold, it is considered that a magnetic anomaly is detected at the corresponding position.

[0021] The single vector magnetometer motion direction magnetic gradient anomaly detection system is used for implementing the single vector magnetometer motion direction magnetic gradient anomaly detection method, and the system comprises:

[0022] The signal acquisition module is used for acquiring the x-axis motion direction magnetic gradient signal through spatial sampling based on a single vector magnetometer cooperating with a stabilizing device.

[0023] The preprocessing module is used for preprocessing the x-axis motion direction magnetic gradient signal.

[0024] The detection module is used for constructing a detector based on the preset orthogonal base function of the x-axis motion direction magnetic gradient, inputting the signal after preprocessing into the detector, and performing magnetic gradient anomaly detection.

[0025] Optionally, the signal acquisition module comprises:

[0026] The single vector magnetometer sensor is used for acquiring the x-axis motion direction magnetic gradient signal through spatial sampling.

[0027] The stabilizing device is used for stabilizing the vector magnetometer sensor.

[0028] Optionally, in the detection module, the preset orthogonal base function is:

[0029]

[0030] wherein, represents the preset orthogonal base function, and ω represents the ratio of the X-axis direction distance to the CPA distance.

[0031] The single vector magnetometer motion direction magnetic gradient anomaly detection system has the following beneficial effects:

[0032] The present application is directed to the problem that the existing vector magnetic gradient anomaly detection system is complicated to correct and has insufficient detection performance under low signal-to-noise ratio, and proposes a single vector magnetometer motion direction magnetic gradient anomaly detection method based on OBF. Firstly, the static and dynamic error characteristics of the vector magnetometer are analyzed; the energy and frequency domain characteristics of the z / y direction and x direction magnetic gradient basis function decomposition are studied, and through comparative analysis, it is found that the x direction magnetic gradient orthogonal basis function e4 has overwhelming advantages in suppressing high frequency noise and detecting weak magnetic signals, and a new OBF detector is designed accordingly; the selection principle of the spatial sampling distance is studied, and the influence of platform jitter and device error on the detection performance of the present application and the traditional OBF detector is analyzed. Simulation experiments show that the single vector magnetometer x direction gradient OBF detector proposed in the present application has lower false spectral peak and more accurate detection than the z / y direction double vector magnetometer array gradient OBF detector, and can effectively overcome the influence of device error on the detection performance. BRIEF DESCRIPTION OF DRAWINGS

[0033] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description only constitute some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0034] Figure 1 Ideal and error coordinate system of the vector magnetometer in the embodiments of the present application;

[0035] Figure 2 Carrier coordinate system in the embodiments of the present application;

[0036] Figure 3 Relative motion relationship between the target and the magnetic sensor in the embodiments of the present application;

[0037] Figure 4 Comparison of z / y and x direction scalar magnetic gradient decomposition basis function frequency spectrum in the embodiments of the present application; wherein (a) is the x direction scalar magnetic gradient decomposition basis function normalized frequency spectrum, and (b) is the y direction scalar magnetic gradient decomposition basis function normalized frequency spectrum;

[0038] Figure 5 Comparison of x direction gradient and z / y direction gradient OBF detector in the embodiments of the present application;

[0039] Figure 6The influence of the error coefficient moment and sway of the vector magnetometer in this embodiment of the invention on the performance of the x-direction gradient OBF detector and the z / y-direction gradient OBF detector; wherein, (a) is the detection performance when the error coefficient matrices of vector magnetometers 1 and 2 are the same and the sway amplitude is 1°, and (b) is the detection performance when the error coefficient matrices of vector magnetometers 1 and 2 are different and the sway amplitude is 0.1°.

[0040] Figure 7 This is a schematic diagram illustrating the actual measurement verification of an embodiment of the present invention;

[0041] Figure 8 This is a schematic diagram of the actual measured output of the target in an embodiment of the present invention; wherein, (a) is the magnetic target signal collected by the system, and (b) is a comparison of the output of the OBF detector in this embodiment with that of the traditional scalar OBF detector;

[0042] Figure 9 This is a schematic diagram of the method for detecting magnetic gradient anomalies in the motion direction of a single-vector magnetometer according to an embodiment of the present invention. Detailed Implementation

[0043] 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.

[0044] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0045] Example 1

[0046] like Figure 9 As shown, this embodiment proposes a method for detecting magnetic gradient anomalies in the motion direction of a single-vector magnetometer, including:

[0047] Based on single-sensor magnetic gradient measurement technology, magnetic gradient signals in the x-axis motion direction are acquired through spatial sampling.

[0048] The magnetic gradient signal in the x-axis motion direction is preprocessed;

[0049] Preset orthogonal basis functions for the magnetic gradient in the x-axis motion direction;

[0050] Based on preset orthogonal basis functions, magnetic gradient anomaly detection is performed on the preprocessed signal.

[0051] Furthermore, the magnetic gradient signal in the x-axis motion direction is:

[0052]

[0053] Wherein, G(x) represents a magnetic gradient signal, B(x+Δx) represents a magnetic induction intensity at x+Δx, B(x) represents a magnetic induction intensity at x, and Δx represents a fixed interval distance.

[0054] Further, the pre-processing of the x-axis motion direction magnetic gradient signal comprises:

[0055] High-frequency noise is suppressed by band-pass filtering.

[0056] Further, the preset orthogonal basis function is:

[0057]

[0058] Wherein, represents a preset orthogonal basis function, and ω represents a ratio of an X-axis direction distance to a CPA (closest point of approach) distance.

[0059] Further, the magnetic gradient anomaly detection comprises:

[0060] The matching degree of the signal after pre-processing and the preset orthogonal basis function is calculated, and if the matching result exceeds a set threshold, it is considered that a magnetic anomaly is detected at the corresponding position.

[0061] Specifically, the motion direction magnetic gradient signal decomposition basis function is derived in the embodiment; by analyzing the influence of various errors of the vector magnetometer on the gradient measurement, and combining the basis function energy and the frequency domain characteristics, a single vector magnetometer motion direction magnetic gradient anomaly detection method based on a stable platform is proposed.

[0062] The embodiment first analyzes various error sources of the vector magnetometer magnetic field measurement and establishes a comprehensive error model; on this basis, a single vector magnetometer motion direction magnetic gradient anomaly OBF detector is researched and designed, and the influence of the geomagnetic background field, platform shaking, and space sampling interval on the performance thereof is analyzed. Simulation experiments show that the method of the embodiment is superior to the traditional vertical direction array double-sensor column magnetic gradient anomaly OBF detection method in low noise ratio and platform shaking conditions, and the real measurement experiment based on the MEMS vector magnetometer further verifies the magnetic anomaly detection effect thereof. The embodiment has the advantages of strong real-time performance, low cost, and small size, and shows good application prospects in dangerous exploration and explosion disposal, intrusion monitoring, and the like.

[0063] Specifically, the vector magnetometer error model is first constructed in the embodiment;

[0064] The measurement accuracy of magnetometer determines the highest level of measurement that the magnetic detection system can achieve. MEMS vector magnetometer magnetic measurement system has the advantages of small size, low power consumption and low cost, but the disadvantage is that the measurement error is large. The error mainly comes from the design and manufacturing defects of the sensor device and the interference of the external environment, which can be divided into static error and dynamic carrier error.

[0065] 1.1 Static error;

[0066] (1) Hard and soft magnetic interference:

[0067] Hard magnetic interference mainly refers to the fixed magnetic field generated by the magnetization of the hard magnetic material of the carrier, which can be regarded as a constant. Soft magnetic interference is generated by the magnetization of soft magnetic materials with small coercivity on the carrier by the geomagnetic field or electromagnetic field. Its size and direction relative to the system are related to the number and mass of soft magnetic materials, the relative position of the magnetic sensor and the carrier, the working state of the surrounding components, and the heading of the system. Let the geomagnetic vector be represented in the magnetometer coordinate system as The measured value of the sensor affected by soft and hard magnetic effects is According to the Poisson model, we have:

[0068] (1)

[0069] In the formula, the matrix is the soft magnetic interference matrix, is the hard magnetic interference.

[0070] (2) Non-orthogonal error:

[0071] Non-orthogonal error is the deviation of the three sensitive axes of the vector magnetometer from the ideal perpendicular state. Let the ideal orthogonal coordinate system of the magnetic sensor be OXYZ, and the actual three-axis non-orthogonal coordinate system of the sensor be OX1Y1Z1. Assume that the Z1 axis coincides with the Z axis; the Y1OZ1 coordinate plane is coplanar with the YOZ plane, and assume that the angle between the OY axis and the OY1 axis is α, the angle between the OX axis and the OX1 axis is γ, and the projection of the OX1 axis on the YOZ plane is the angle β with the OY axis, as shown in Figure 1 The output of the vector magnetometer is:

[0072] (2)

[0073] In the formula, α, β, and γ are parameters that describe the non-orthogonality of the three axes of the vector magnetometer.

[0074] (3) Sensitivity error and zero offset error:

[0075] The inter-axis static sensitivity error is a measurement error caused by different sensitivities of each axis of the vector magnetometer and different characteristics of the amplification circuit of the measurement signal. Assuming that the sensitivities of the three axes of the magnetic sensor are kx, ky, and kz, considering the residual magnetism of the core of the vector magnetometer and the zero-point drift of the analog circuit, assuming that the residual magnetism and the zero-point drift of the circuit are constant values, denoted as The mathematical model of the sensitivity error and the zero-point offset error is:

[0076] (3)

[0077] In the formula, [B x2 B y2 B z2 ] is the three-dimensional magnetic field strength under the condition of the existence of the sensitivity error and the zero-point offset error.

[0078] 1.2 Dynamic error;

[0079] The moving carrier will generate a dynamic magnetic interference field. At present, the carrier magnetic compensation technology is mainly based on the T-L model proposed by Tolles and Lawson. The model divides the carrier magnetic interference source into three parts: constant field, induced field, and eddy current field. The constant field refers to the magnetic field generated by the hard magnetic material in the carrier platform, which is fixed in the carrier coordinate system. The induced field refers to the magnetic field generated by the soft magnetic material in the carrier platform being magnetized by the geomagnetic field. The eddy current field refers to the magnetic field generated by the eddy current when the aircraft moves in the geomagnetic field, which is proportional to the acceleration of the carrier maneuver.

[0080] Considering the following unmanned aerial vehicle (UAV) airborne magnetic platform, an aircraft carrier coordinate system is established as shown in Figure 2 . The coordinate origin is located at the center of mass of the aircraft, the X-axis is positive in the left direction along the flight direction, the Y-axis is positive in the forward direction of the aircraft, and the Z-axis and the X-axis and the Y-axis form a right-handed coordinate system. The angle between the geomagnetic field and the three axes of the carrier coordinate system is a, θ, and φ, respectively.

[0081] Assuming that the UAV moves at a constant speed in a straight line, the eddy current field is 0, and the carrier magnetic compensation model is represented as:

[0082] (4)

[0083] where, is the total magnetic interference field, is the constant field, is the induced field. The constant field can be regarded as a constant vector in the carrier coordinate system, that is:

[0084] (5)

[0085] The size of the induction field is proportional to the applied magnetic field. In the carrier coordinate system, the size and direction of the induction field change with the carrier attitude, and its expression is:

[0086] (6)

[0087] where, is the total magnetic interference field, a 11 ~a 33 is the coefficient affected by soft magnetic material, which represents the physical meaning of the induced magnetic field generated by the projection of the geomagnetic field on the carrier coordinate system. Since the dynamic error is an environmental disturbance, gradient measurement can be used to suppress it.

[0088] 1.3 Comprehensive error modeling

[0089] Based on the above different source errors, the mathematical model of the vector magnetometer error in the actual carrier environment is established:

[0090] (7)

[0091] In the formula: represents the three-axis non-orthogonal error matrix.

[0092]

[0093] represents the soft iron interference matrix. represents the direction cosine matrix corresponding to the carrier attitude angle; represents the mapping of the local geomagnetic vector in the navigation coordinate system, and the initial value of the vector can be obtained by querying the international geomagnetic vector website; represents the hard iron interference; represents the measurement noise. The navigation coordinate system is defined as the East North Sky coordinate system, and the carrier coordinate system is defined as the right front up coordinate system. Let:

[0094] , ,

[0095] Then formula (2) can be expressed as:

[0096] (8)

[0097] In the same environment, the parameters C and in formula (8) can be considered as unknown constant parameters.

[0098] Specifically, the inference steps of the orthogonal basis function (OBF) based magnetic gradient anomaly detection in the present embodiment are as follows:

[0099] 2.1. z / y direction magnetic gradient orthogonal basis decomposition:

[0100] When the detection distance is greater than 2.5 times the length of the magnetic target, the magnetic target can be regarded as a magnetic dipole, and the magnetic field vector B at a distance r from the magnetic target can be expressed as:

[0101] (9)

[0102] In the formula, μ0= 4π×10 －7 H / m is the space permeability; M represents the target magnetic moment vector; r represents the modulus of the displacement vector r of the magnetic target to the sensor. When (T represents the geomagnetic field vector), the scalar magnetic field S = |B| is recorded, and the measurement of S can be regarded as the projection of the magnetic field vector in the direction of the geomagnetic field, that is:

[0103] (10)

[0104] In the formula, γ is the angle between the target magnetic field vector B and the geomagnetic vector.

[0105] Since the magnetic anomaly signal is very weak compared to the earth background noise, slight shaking of the carrier can annihilate the anomaly signal, and a multi-sensor array gradient measurement method can be used to remove the background magnetic field and improve the detection performance. The existing technology proposes a y / z direction double scalar sensor magnetic field gradient OBF detector, in which the z direction magnetic gradient sensor is composed of two magnetic sensors with a distance of Δz, as shown in Figure 3 . Among them, the coordinate origin O represents the position of the target (dipole), the X axis represents the direction of the linear motion of the moving carrier, the Y axis represents the port side direction of the moving carrier, and the Z axis represents the vertical upward direction; R0 represents the closest point of approach (CPA) distance between the sensor and the magnetic target; D0 represents the distance between the current position of the sensor and the CPA point, θ M represents the angle between the target magnetic moment direction and the Z axis, θ T represents the angle between the geomagnetic field direction and the Z axis, Φ M represents the angle between the projection of the target magnetic moment vector in the OXY plane and the X axis, and Φ T represents the angle between the projection of the geomagnetic field direction in the OXY plane and the X axis.

[0106] The scalar magnetic gradient measurement G is equivalent to the difference of the projection of the magnetic field vector in the direction of the geomagnetic field, and the differential gradient is equal to the theoretical gradient value plus the high-order term in the following part. Because the high-order derivative of the magnetic field decays rapidly with distance, the magnetic field difference result can be understood as an approximate result of the gradient theoretical value. Taking the differential of formula (10) with respect to z, the base function expression can be derived as:

[0107] (11)

[0108] wherein: is a constant term, M is the modulus of the target magnetic moment vector M; ω is the ratio of the X-axis direction distance to the CPA distance, ω =D0 / R0; C j are the basis function coefficients, φ j (w) is the magnetic gradient basis function. Orthogonalization of the basis function in the above decomposition can be expressed as follows:

[0109] (12)

[0110] wherein: v j (w) is the orthogonalization function of φ j (w).

[0111] (13)

[0112] The scalar magnetic gradient calculation along the y direction has symmetry with the z-axis direction, and the orthogonal basis function form is the same.

[0113] 2.2 Based on the direction of motion magnetic gradient anomaly detection:

[0114] The above y / z direction scalar magnetic gradient detection scheme requires the deployment of two sensors. If the sensor is arranged along the x-axis direction, since it is consistent with the direction of motion, the sampling points in the direction of motion can be used to calculate the horizontal gradient, thereby reducing one scalar sensor and simplifying the system arrangement. Taking the differential of formula (10) with respect to the x direction, the horizontal direction magnetic gradient orthogonal basis representation form can be derived as:

[0115] (14)

[0116] wherein, the basis function ψ j (w) and the corresponding orthogonalization function e j (w) are:

[0117] (15)

[0118] It can be seen that the x direction horizontal gradient basis function is different from the z / y direction basis function.

[0119] The traditional anomaly detection method based on orthogonal basis function usually uses each orthogonal basis function to multiply and integrate with the signal to be detected respectively, and obtains the energy detector to construct the energy detector for anomaly detection by squaring the weight of the orthogonal basis function. This method only uses the orthogonality of the orthogonal basis function, but does not consider the energy distribution characteristics of the basis function itself and the signal characteristics of the magnetic gradient. Table 1 and Figure 4 respectively give the signal energy and frequency domain graph of the z / y and x direction scalar magnetic gradient decomposition basis function.

[0120] Table 1 z / y and x direction scalar magnetic gradient decomposition basis function energy contrast

[0121]

[0122] It can be seen that the z / y direction gradient decomposition basis function signal energy is the highest φ3, the x direction magnetic gradient decomposition basis function has the highest signal energy component ψ4, φ3 and φ3 are both even functions and the signal energy is mainly concentrated in low frequency, which is suitable for matching steady state signal, and compared with φ3, the energy of φ3 is more concentrated in low frequency. Since the noise energy is mainly in high frequency, it can be inferred that the x direction magnetic gradient orthogonal basis function e4 is a natural high frequency noise suppression for the detector, and the detection effect of weak magnetic signal should be better than the y direction magnetic gradient orthogonal basis function v3 as the detector and the traditional all orthogonal basis function weight square sum detector.

[0123] 2.3 Analysis of the influence of spatial sampling distance on detection performance:

[0124] The z / y and x direction gradient measurement are both a kind of spatial sampling. According to the Nyquist sampling theorem, if the array or spatial sampling frequency is lower than twice the signal frequency, it may lead to distortion of the gradient measurement signal. For example, the frequency range of the orthogonal basis function e4 is about [-1, 1], so the sampling distance dx should satisfy dx < 1 / 2 meter. dw4

[0125] Secondly, the measurement noise is also differentiated in the gradient calculation process. Assuming that the noise variance of each sampling point is σ n , then the noise σ g in the gradient signal can be calculated as follows:

[0126] 16)

[0127] It can be seen that with the decrease of dx, the noise variance in the sampling sequence is increasing.

[0128] On the other hand, the gradient basis function decomposition is calculated by difference to calculate the magnetic field differential, and the spatial sampling distance is too large, which will lead to the increase of the error of magnetic field differential calculation.

[0129] In summary, the spatial sampling distance is selected as 0.4-1.3 meters.

[0130] 2.4 Analysis of the influence of shaking and device error on detection performance:

[0131] ​When the dual vector magnetometer array is used to collect magnetic vector data to calculate the vector magnetic field gradient, the magnetic field gradient measurement is not only affected by the platform shaking, but also affected by the sensor inherent error. When the dual vector magnetometer is used to calculate the gradient, the magnetic fields at the two sensors are B1 and B2 respectively, the background magnetic field is T, and the error coefficient matrices of the two sensors are K1 and K2 respectively; according to formula (8), the carrier shaking changes the Euler attitude matrix C n , then the outputs of the two sensors are:

[0132] (17)

[0133] Obviously, due to the difference of the error coefficient matrix of the vector magnetometer, it is impossible to completely offset the background magnetic field by using two vector magnetometers to form an array whether measuring the total field gradient or the vector gradient, and the measurement result will still be disturbed, and when using a single vector magnetometer to calculate the gradient, the fixed errors such as hard magnetic interference and zero drift and the dynamic interference errors of the carrier can be completely offset, but the attitude needs to be kept unchanged. At the same time, the induced magnetic field caused by the carrier motion is related to the attitude of the carrier, and the attitude of the carrier needs to be kept stable and unchanged at each measuring point to offset the induced magnetic field as much as possible.

[0134] The following embodiment is verified by experiment:

[0135] 3.1 x-direction gradient and z / y-direction gradient OBF detector comparison simulation:

[0136] Suppose the center of the magnetic target is the origin of the spatial rectangular coordinate system, the X-axis coincides with the axis of the platform and points to the bow of the platform, the Y-axis points to the port side, and the Z-axis is vertically upward. The simulation conditions are set as follows: the initial position of the target is [0, 1.2, 0.8] meters. The target magnetic moment is (3A·m 2 , 0.2A·m 2 , 0.3A·m 2 ), the horizontal motion direction gradient tensor system measurement space sampling length is 1.1m, and the vertical gradient sensor aperture is 0.5m; the observation window interval is [-2R0, 2R0]. Figure 5 For the signal-to-noise ratio of-20dB, the detection performance comparison of the horizontal motion direction gradient OBF and the vertical gradient OBF detector shows that the detection curve of the horizontal motion direction gradient OBF has a low false spectral peak, and has more accurate detection performance.

[0137] 3.2 Simulation analysis of the influence of carrier platform shaking and device error:

[0138] The influence of the shaking of the carrier platform and the error of different vector magnetometers on the performance of the horizontal motion direction gradient and the vertical gradient OBF detector is considered. The error coefficient matrices of the vector magnetometers 1 and 2 are respectively:

[0139]

[0140] Bias error is b1=[35, 47, 29]nT, b2=[40, 50, 30]nT. Figure 6 (a) in FIG. 7 is that the error coefficient matrix of the vector magnetometer 1 and the vector magnetometer 2 is assumed to be the same, and the signal-to-noise ratio is -12.5 dB. At this time, even when the platform swing amplitude is 1°, the performance of the x-direction gradient OBF detector and the z / y-direction gradient OBF detector does not change significantly. Figure 6 (b) in FIG. 7 is that the vector magnetometer 1 and the vector magnetometer 2 measure the gradient according to the respective error coefficient matrices, the signal-to-noise ratio is -12.5 dB, and the swing amplitude is 0.1. At this time, it can be seen that the z-direction gradient OBF detector has completely failed to detect, and it can be seen that the error of the vector magnetometer device has a decisive influence on the detection performance.

[0141] 3.3 Actual measurement experiment:

[0142] FIG. 7 shows a detection experiment performed on a playground of a certain university. The magnetic gradient measurement system is a MEMS vector magnetometer installed on a DJI handheld gimbal such as a shadow RS4mini platform. The measurement error mean square is 200 nT. The target to be detected is a stationary cylindrical permanent magnet (radius 2.5 mm, height 30 mm), with coordinates [0, 1.5, 1.0] meters, and the estimated magnetic moment is (3A·m 2 , 0.2A·m 2 , 0.3A·m 2 ). The DJI shadow RS4 handheld stabilizer is used as the magnetometer installation platform, and the angle stability can reach 0.04°.

[0143] Figure 8 (a) in FIG. 8 is the magnetic target signal collected by the system, Figure 8 (b) in FIG. 8 is the abnormal detection curve displayed by using the algorithm of the embodiment. It can be seen that compared with the traditional scalar OBF detector, the OBF detector proposed in the embodiment has more accurate detection performance on weak magnetic target signals. If a high-performance fluxgate sensor is used, such as a measurement error of 0.1 nT, and a magnetic moment of 100 A·m 2 order of magnitude is considered, the measurement distance can be more than 40 m.

[0144] The embodiment is directed to the problem that the existing vector magnetic gradient anomaly detection system is complicated to correct and has insufficient detection performance under low signal-to-noise ratio, and proposes a single vector magnetometer motion direction magnetic gradient anomaly detection method based on OBF. Firstly, the static and dynamic error characteristics of the vector magnetometer are analyzed; the energy and frequency domain characteristics of the z / y direction and x direction magnetic gradient basis function decomposition are studied, and through comparative analysis, it is found that the x direction magnetic gradient orthogonal basis function e4 has overwhelming advantages in suppressing high frequency noise and detecting weak magnetic signals, and accordingly a new OBF detector is designed; the space sampling distance selection principle is studied, and the influence of platform jitter and device error on the detection performance of the embodiment and the traditional OBF detector is analyzed. Simulation experiments show that the single vector magnetometer x direction gradient OBF detector proposed in the embodiment has lower false spectral peaks and more accurate detection than the z / y direction double vector magnetometer array gradient OBF detector, and can effectively overcome the influence of device error on the detection performance. The actual measurement experiment further verifies that the method of the embodiment has good detection effect on weak magnetic target signals under low signal-to-noise ratio and platform jitter interference, and discusses the potential of the method for improving the detection performance of the high-precision fluxgate magnetic gradient anomaly detection system.

[0145] Embodiment 2

[0146] The embodiment proposes a single vector magnetometer motion direction magnetic gradient anomaly detection system, which comprises:

[0147] A signal acquisition module is configured to acquire x-axis motion direction magnetic gradient signals through spatial sampling based on a single vector magnetometer and a stabilizing device.

[0148] A preprocessing module is configured to preprocess the x-axis motion direction magnetic gradient signals.

[0149] A detection module is configured to construct a detector based on a preset orthogonal basis function of the x-axis motion direction magnetic gradient, input the preprocessed signals into the detector, and perform magnetic gradient anomaly detection.

[0150] Further, the signal acquisition module comprises:

[0151] A single vector magnetic intensity sensor is configured to acquire x-axis motion direction magnetic gradient signals through spatial sampling.

[0152] A stabilizing device is configured to stabilize the vector magnetic intensity sensor.

[0153] Specifically, in the embodiment, the sensor deployment is as follows: a vector magnetometer (such as PNI RM3100, a three-axis fluxgate sensor) is selected; the sensor is installed on a three-axis stabilizing gimbal to ensure that the attitude error is less than 0.05°.

[0154] Data acquisition: vector magnetometer with stable gimbal (such as DJI RS4mini, stability accuracy 0.04°) through space sampling to calculate the motion direction (x-axis) magnetic gradient (sampling interval 0.4-1.3 meters, preferably 1.1 meters); uniform motion (speed v=1m / s) at fixed intervals (Δt=1.1s) sampling. Real-time gradient calculation: G(x)=[B(x+Δx)-B(x)] / Δx.

[0155] Specifically, in this embodiment, signal preprocessing: band-pass filtering (0.1-5Hz) to suppress high-frequency noise (no need to pre-calibrate matrix C and bias b0).

[0156] Further, in the detection module, the preset orthogonal basis function is:

[0157]

[0158] wherein, ω represents the ratio of the X-axis direction distance to the CPA distance.

[0159] Specifically, in this embodiment, the new OBF detector design: derive the x-direction magnetic gradient orthogonal basis function, select the energy concentrated and low frequency characteristic optimal e4 (energy ratio 85.9%) as the main detection basis function (compared with z / y direction basis function φ3 energy 70.7%).

[0160] Example code (core algorithm) of this embodiment:

[0161] def e4_basis(w):

[0162] return 16 / sqrt(70*pi-7)*(1 / (1+w**2)**2.5 - sqrt(7) / 16*w**2 / (1+w**2)**3.5)

[0163] def detect(signal):

[0164] weights = [sum(signal * e4_basis(w)) for w in window]

[0165] return sqrt(sum(w**2 for w in weights))

[0166] Parameter optimization:

[0167] Space sampling interval: balance signal bandwidth, noise amplification and differential error, preferably 0.5-1.2 meters.

[0168] Observation window: [-2R0, 2R0], R0 is set by prior estimation;

[0169] Scenario: University playground, target magnetic moment 3A·m² (as Figure 7

[0170] Results: At a signal-to-noise ratio of -15 dB, the detection probability reached 92% (traditional method < 65%).

[0171] This embodiment first combines the single vector magnetometer motion direction gradient with the specific base function e4; combined with the actual measurement, it is speculated that the algorithm can achieve a detection distance of 40 m on a 200 nT level MEMS device (when the magnetic moment is 100 A·m²).

[0172] The above-described embodiments only describe the preferred modes of the present application and do not limit the scope of the present application. Without departing from the design spirit of the present application, various modifications and improvements to the technical solutions of the present application made by those skilled in the art shall fall within the protection scope determined by the claims of the present application.​

Claims

1. A method of detecting magnetic gradient anomalies in the direction of motion using a single vector magnetometer, characterized in that, The method comprises the following steps: Based on single sensor magnetic gradient measurement technology, the x-axis motion direction magnetic gradient signal is obtained by spatial sampling; The x-axis motion direction magnetic gradient signal is preprocessed; The orthogonal basis function of the x-axis motion direction magnetic gradient is preset; The preset orthogonal basis function is: wherein, wherein, denotes a preset orthogonal basis function, and ω denotes a ratio of the distance in the x-axis direction to the distance of the closest approach point. Based on the preset orthogonal basis function, the magnetic gradient anomaly detection is performed on the preprocessed signal.

2. The single vector magnetometer motion direction magnetic gradient anomaly detection method of claim 1, wherein, The x-axis motion direction magnetic gradient signal is: G(x)=[B(x+Δx)-B(x)] / Δx Wherein, G(x) represents the magnetic gradient signal, B(x+Δx) represents the magnetic induction intensity at x+Δx, B(x) represents the magnetic induction intensity at x, and Δx represents a fixed interval distance.

3. The single vector magnetometer motion direction magnetic gradient anomaly detection method of claim 1, wherein, The preprocessing of the x-axis motion direction magnetic gradient signal comprises: High-frequency noise is suppressed by band-pass filtering.

4. The single vector magnetometer motion direction magnetic gradient anomaly detection method of claim 1, wherein, The magnetic gradient anomaly detection comprises: The matching degree of the signal after signal preprocessing and the preset orthogonal basis function is calculated, and if the matching result exceeds the set threshold, it is considered that the magnetic anomaly is detected at the corresponding position.

5. A single vector magnetometer motion direction magnetic gradient anomaly detection system characterized by, The system for implementing the single vector magnetometer motion direction magnetic gradient anomaly detection method according to any one of claims 1-4 comprises: The signal acquisition module is used for acquiring the x-axis motion direction magnetic gradient signal by spatial sampling based on a single vector magnetometer cooperating with a stabilizing device; The preprocessing module is used for preprocessing the x-axis motion direction magnetic gradient signal; The detection module is used for constructing a detector based on the preset orthogonal basis function of the x-axis motion direction magnetic gradient, inputting the preprocessed signal into the detector, and performing magnetic gradient anomaly detection; In the detection module, the preset orthogonal basis function is: wherein, represents a preset orthogonal basis function, and ω represents a ratio of the distance in the X-axis direction to the distance of the closest approach point.

6. The single vector magnetometer motion direction magnetic gradient anomaly detection system of claim 5, wherein, The signal acquisition module comprises: A single vector magnetic intensity sensor is used for acquiring the x-axis motion direction magnetic gradient signal by spatial sampling; The stabilizing device is used for stabilizing the vector magnetic intensity sensor.

Images