Method and test system for coupling a magnetic gradiometer to a squid to improve sensitivity

Abstract

The application relates to the technical field of magnetic detection, in particular to a method and a test system for coupling a magnetic gradiometer and a SQUID to improve sensitivity, which comprises the following steps: deploying at least one set of magnetic gradiometer probe arrays in a superconducting shielding environment to obtain original magnetic gradiometer tensor data streams of a target area; feeding a multi-channel SQUID front-end input coil through a superconducting transmission line; completing magnetic flux locking and amplification to generate a preliminary signal by the SQUID; suppressing environmental and SQUID background noise by using an adaptive digital noise reduction processor to obtain a high-fidelity magnetic gradiometer signal; and generating a high-resolution magnetic gradiometer distribution map through gradient reconstruction and spatial resolution enhancement. The superconducting shielding environment combines array acquisition to suppress external interference and retain complete tensor characteristics, and the superconducting transmission line combines a multi-channel SQUID to reduce transmission noise and cumulative error, so that the signal is closer to the original state, and the SQUID locking sensitivity and amplification fidelity are improved. The application improves the sensitivity of magnetic gradiometer detection and the accuracy of target identification.

Description

Technical Field

[0001] This invention relates to the field of magnetic detection technology, and more particularly to a method and testing system for coupling a magnetic gradiometer with a SQUID to improve sensitivity. Background Technology

[0002] In weak magnetic target detection scenarios, existing magnetic gradient measurement techniques often employ conventional magnetic gradiometers or single-channel SQUID systems to acquire magnetic signals. However, these methods are susceptible to interference from stray magnetic fields and transmission noise, making it difficult to reliably extract high-fidelity magnetic gradient information. To obtain magnetic gradient tensor data, conventional solutions typically deploy a single probe or a small number of dispersed probes in an open or semi-shielded environment to directly acquire the magnetic gradient signal of the target area. The signal is then transmitted to a SQUID for processing via a standard metal transmission line. This approach lacks effective isolation from low-frequency magnetic fields and high-frequency electromagnetic interference, and the probe layout does not provide array-based tensor acquisition capabilities. Consequently, the signal-to-noise ratio of the raw data is limited, making it difficult to reflect the spatial tensor characteristics of the target's magnetic field.

[0003] In conventional solutions, the magnetic gradiometer probes are not integrated into an array within a superconducting shielded environment, making it impossible to simultaneously suppress external static magnetic field fluctuations and high-frequency radiated noise. The acquired raw magnetic gradient tensor data stream is susceptible to environmental interference and distortion. Furthermore, the acquired signal is fed into the SQUID front end via a conventional conductive transmission line, where Joule thermal noise and electromagnetic induction noise caused by the transmission line resistance are superimposed on the signal, reducing the SQUID's response accuracy to weak magnetic flux. The solution lies in addressing the issues of deploying at least one magnetic gradiometer probe array within a superconducting shielded environment to obtain the raw magnetic gradient tensor data stream, and in feeding the raw magnetic gradient tensor data stream into the multi-channel SQUID front end input coil via a superconducting transmission line. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and to propose a method and testing system for improving sensitivity by coupling a magnetic gradient meter with a SQUID.

[0005] To achieve the above objectives, the present invention employs the following technical solution: a method for coupling a magnetic gradient meter with a SQUID to improve sensitivity, comprising:

[0006] Deploy at least one set of magnetic gradient meter probe arrays in a superconducting shielded environment to acquire the original magnetic gradient tensor data stream of the target region;

[0007] The original magnetic gradient tensor data stream is fed into the multi-channel SQUID front-end input coil through a superconducting transmission line.

[0008] Magnetic flux locking and signal amplification processing are performed in the multi-channel SQUID to generate a magnetic gradient signal waveform that has been preliminarily amplified.

[0009] The magnetic gradient signal waveform is subjected to joint suppression of environmental noise and SQUID background noise using a pre-set adaptive digital noise reduction processor to generate a noise-reduced high-fidelity magnetic gradient signal.

[0010] The high-fidelity magnetic gradient signal is subjected to gradient reconstruction and spatial resolution enhancement calculations to generate a high-resolution magnetic gradient distribution map for target recognition.

[0011] As a further aspect of the present invention, the step of deploying at least one set of magnetic gradient meter probe arrays in a superconducting shielded environment to acquire the original magnetic gradient tensor data stream of the target region includes:

[0012] A vacuum cryogenic cavity was constructed inside a superconducting shield, and a set of magnetic gradient meter probe arrays was immersed in the liquid helium environment of the vacuum cryogenic cavity.

[0013] The magnetic gradiometer probe array consists of three orthogonally placed basic magnetic gradiometer units, each of which contains a pair of spatially fixed superconducting induction coils.

[0014] The pair of superconducting induction coils in each basic magnetic gradient meter unit are driven to work synchronously to measure the difference in magnetic flux intensity between two points in space in differential mode.

[0015] Each basic magnetic gradiometer unit outputs in real time the raw data sequence of the difference in magnetic flux density it senses as a function of time;

[0016] The raw data sequences output by the three orthogonal directions of the basic magnetic gradient meter unit are aligned in time and packaged to form the raw magnetic gradient tensor data stream containing the gradient data of the three components.

[0017] As a further aspect of the present invention, the step of feeding the original magnetic gradient tensor data stream into the multi-channel SQUID front-end input coil via a superconducting transmission line includes:

[0018] Configure a superconducting transmission line pair that corresponds one-to-one with each basic magnetic gradiometer unit in the magnetic gradiometer probe array;

[0019] The superconducting leads at the output of each basic magnetic gradiometer unit are connected to the positive and negative input ends of the corresponding superconducting transmission line pair to form a differential signal transmission link.

[0020] The output of each pair of superconducting transmission lines is connected to the input of an independent input channel in a multi-channel SQUID, the input of which includes a set of planar helical input coils for flux coupling.

[0021] The mutual inductance coefficient between the planar spiral input coil and the SQUID loop is kept stable within a preset range to ensure effective coupling of the magnetic gradient signal.

[0022] As a further aspect of the present invention, the step of performing magnetic flux locking and signal amplification processing in the multi-channel SQUID to generate a pre-amplified magnetic gradient signal waveform includes:

[0023] A preset bias current is applied to the SQUID loop of each SQUID input channel to make it operate in the linear region of the flux-voltage conversion curve;

[0024] The output voltage signal of the multi-channel SQUID is converted into a feedback current through a feedback resistor;

[0025] The feedback current drives a tightly coupled feedback coil to generate a compensating magnetic flux, which counteracts the magnetic flux change introduced by the input coil, thus forming a magnetic flux locking loop.

[0026] The feedback current value of each channel is collected in real time, and the feedback current value is proportional to the change in input magnetic flux;

[0027] The feedback current signal of each channel is converted into a voltage signal by a transimpedance amplifier. The amplitude of the voltage signal is amplified to form the pre-amplified magnetic gradient signal waveform.

[0028] As a further aspect of the present invention, the step of using a pre-set adaptive digital noise reduction processor to jointly suppress environmental noise and SQUID background noise on the magnetic gradient signal waveform to generate a noise-reduced high-fidelity magnetic gradient signal includes:

[0029] Acquire reference noise signal waveforms generated solely by the environment and the system itself in a targetless state;

[0030] Establish a cross-correlation function model between the noise components in the reference noise signal waveform and the real-time magnetic gradient signal waveform;

[0031] Based on the cross-correlation function model, the estimated noise component is adaptively subtracted from the real-time magnetic gradient signal waveform using the minimum mean square error algorithm.

[0032] To filter out the periodic spike noise pattern unique to the SQUID background noise, a pattern matching and interpolation replacement algorithm is used.

[0033] Wavelet transform threshold denoising is performed on the processed signal waveform to further smooth the random noise floor.

[0034] The final output is a high-fidelity magnetic gradient signal with significantly improved signal-to-noise ratio in both the frequency and time domains.

[0035] As a further aspect of the present invention, the step of performing gradient reconstruction and spatial resolution enhancement calculations on the high-fidelity magnetic gradient signal to generate a high-resolution magnetic gradient distribution map for target recognition includes:

[0036] The high-fidelity magnetic gradient signals in three orthogonal directions are registered according to the spatial geometric position of the magnetic gradient meter probe array;

[0037] Using the known mathematical relationship between the magnetic gradient tensor and the magnetic field vector, the three-dimensional magnetic field vector distribution at each point in space is calculated by numerical integration algorithm.

[0038] The calculated three-dimensional magnetic field vector distribution is then subjected to spatial interpolation and meshing.

[0039] Based on gridded magnetic field data, calculate all components of the magnetic gradient tensor at each spatial grid point;

[0040] The calculated magnitude of the magnetic gradient tensor components or the component corresponding to the maximum principal eigenvalue of the magnetic gradient tensor is mapped to a three-dimensional spatial coordinate system, and then rendered to generate three-dimensional contour lines or pseudo-color cloud maps.

[0041] The three-dimensional contour lines or pseudo-color cloud map are projected onto a two-dimensional observation plane, and geographical location information is superimposed to form the high-resolution magnetic gradient distribution map used for target identification.

[0042] As a further aspect of the present invention, the method of driving a pair of superconducting induction coils in each basic magnetic gradient meter unit to operate synchronously and measuring the difference in magnetic flux density between two points in space in differential mode includes:

[0043] Each pair of superconducting induction coils in a basic magnetic gradiometer unit is configured with an independent superconducting quantum interference device as a readout circuit;

[0044] A superconducting constant current source is used to provide a stable bias operating point for each readout circuit;

[0045] The original voltage signals output by the readout circuits of a pair of superconducting induction coils are acquired synchronously.

[0046] The original voltage signal is input to the positive and negative input terminals of the differential amplifier;

[0047] The differential amplifier calculates the difference between the two original voltage signals in real time, and the amplified difference directly corresponds to the difference in magnetic induction intensity between the two spatial points.

[0048] The amplified difference signal is converted from analog to digital and used as the effective output of the basic magnetic gradient meter unit.

[0049] As a further aspect of the present invention, the filtering of periodic spike noise patterns unique to the SQUID background noise using a pattern matching and interpolation replacement algorithm includes:

[0050] Monitor the undenoised magnetic gradient signal, and when an isolated noise pulse event is identified, align and average multiple such events to extract the noise pulse waveform as a template.

[0051] In real-time signal processing, the current signal waveform segment is continuously subjected to sliding cross-correlation calculation with the noise pulse waveform template;

[0052] When the cross-correlation value exceeds the dynamic adaptive threshold, it is determined that the signal waveform segment is contaminated by the periodic spike noise;

[0053] Identify the start and end times of noise pulses in the signal waveform segment;

[0054] The signal data within the identified contaminated time point is removed, and the normal signal data from the time points before and after it is used to reconstruct the signal for the corresponding time period through spline interpolation algorithm, thus achieving noise-free replacement.

[0055] As a further aspect of the present invention, the step of utilizing the known mathematical relationship between the magnetic gradient tensor and the magnetic field vector to invert and calculate the three-dimensional magnetic field vector distribution at each point in space through a numerical integration algorithm includes:

[0056] Choose a spatial point with a known background magnetic field strength as the starting point for integration;

[0057] Based on the measured magnetic gradient tensor data, the change in magnetic field vector between adjacent grid points is calculated step by step along the preset spatial integration path.

[0058] The calculation of the change in the magnetic field vector is based on the definition of the magnetic gradient tensor, that is, the rate of change of the magnetic field vector along a certain direction in space is equal to the projection of the magnetic gradient tensor onto that direction.

[0059] By accumulating the changes in the magnetic field vector along the integration path and combining them with the background magnetic field vector at the integration starting point, the magnetic field vector at each grid point on the integration path is calculated.

[0060] By repeating the calculation process through different integration paths, and then using a least squares optimization algorithm to fuse and correct the calculation results of all paths, a three-dimensional magnetic field vector distribution with consistent performance and minimal error across the entire space is obtained.

[0061] As a further aspect of the present invention, when the processor executes the computer program, it implements the steps of the method for coupling a magnetic gradient meter with a SQUID to improve sensitivity as described in any of the above claims.

[0062] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0063] At least one array of magnetic gradiometer probes is deployed in a superconducting shielded environment to acquire the raw magnetic gradient tensor data stream of the target region. The superconducting shielding environment forms an attenuation barrier against external static magnetic fields and broadband electromagnetic interference, improving the uniformity of the magnetic field in the space where the probe array is located. The arrayed layout simultaneously captures the gradient changes of the target magnetic field in different directions, and the resulting raw magnetic gradient tensor data stream retains the complete tensor characteristics of the spatial distribution of the target magnetic field. Compared with single-probe acquisition in conventional open or semi-shielded environments, this method reduces crosstalk of environmental magnetic field fluctuations to tensor component acquisition, allowing the subsequently processed signal to more closely resemble the tensor shape of the target's true magnetic field.

[0064] The original magnetic gradient tensor data stream is fed into the multi-channel SQUID front-end input coil via a superconducting transmission line. The zero-resistance characteristic of the superconducting transmission line eliminates Joule thermal noise during transmission, and its low response to high-frequency electromagnetic induction in the superconducting state reduces additional noise in the transmission path, resulting in minimal magnetic flux loss during signal transmission. The multi-channel SQUID front-end input coil can receive the tensor component signals output from the array probe in parallel, avoiding timing errors and noise accumulation caused by time-division sampling or single-channel switching. Compared to conventional solutions using ordinary metal transmission lines with single-channel or multi-channel non-superconducting transmission, this method makes the signal entering the SQUID closer to the original tensor state acquired by the probe array, improving the SQUID's locking sensitivity to weak magnetic flux changes and the fidelity of signal amplification. Attached Figure Description

[0065] Figure 1 This is a flowchart of the method for coupling a magnetic gradient meter with a SQUID to improve sensitivity according to the present invention;

[0066] Figure 2 A flowchart illustrating the signal feeding into the multi-channel SQUID front-end input coil;

[0067] Figure 3 A plan view for underground pipeline detection that combines magnetic gradient hotspots with safe detection paths;

[0068] Figure 4 Comparison of wavelet transform threshold noise reduction effects;

[0069] Figure 5 This is a high-resolution magnetic gradient tensor norm distribution diagram at a depth of 6 meters underground, indicating the pipeline's orientation. Detailed Implementation

[0070] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0071] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0072] See Figure 1 The method of coupling a magnetic gradiometer with a SQUID to improve sensitivity is carried out in a superconducting shielded environment. At least one set of magnetic gradiometer probe arrays is deployed to obtain the original magnetic gradient tensor data stream of the target region. This data stream is fed into the front-end input coil of a multi-channel SQUID through a superconducting transmission line. Magnetic flux locking and signal amplification processing are performed in the multi-channel SQUID to generate a magnetic gradient signal waveform that has been initially amplified. A pre-set adaptive digital noise reduction processor is used to jointly suppress environmental noise and SQUID background noise on the signal waveform to generate a noise-reduced high-fidelity magnetic gradient signal. Gradient reconstruction and spatial resolution enhancement calculations are performed on the high-fidelity magnetic gradient signal to generate a high-resolution magnetic gradient distribution map for target recognition.

[0073] In one embodiment of the invention, a vacuum cryogenic cavity is constructed inside a superconducting shield. An array of magnetic gradiometer probes is immersed in the liquid helium environment of this cavity. The magnetic gradiometer probe array consists of three orthogonally placed basic magnetic gradiometer units. Each basic magnetic gradiometer unit includes a pair of spatially fixed superconducting induction coils. The pair of superconducting induction coils in each basic magnetic gradiometer unit are driven to operate synchronously, measuring the difference in magnetic flux density between two points in space in differential mode. An independent superconducting quantum interference device (SQU) is configured as a readout circuit for each pair of superconducting induction coils in each basic magnetic gradiometer unit. A superconducting constant current source provides a stable bias operating point for each readout circuit. The original voltage signals output from the readout circuits of a pair of superconducting induction coils are acquired in one step. These original voltage signals are then input to the positive and negative input terminals of a differential amplifier. The differential amplifier calculates the difference between the two original voltage signals in real time. This difference, after amplification, directly corresponds to the difference in magnetic flux density between two spatial points. The amplified difference signal is then converted from analog to digital and used as the effective output of the basic magnetic gradiometer unit. Each basic magnetic gradiometer unit outputs the original data sequence of the magnetic flux density difference it senses over time in real time. The original data sequences output by the three orthogonal basic magnetic gradiometer units are aligned in time and packaged to form an original magnetic gradient tensor data stream containing three component gradient data.

[0074] In practice, a magnetic gradiometer coupled with a SQUID system is configured for highly sensitive, non-destructive detection of buried ferrous pipelines. A vacuum cryogenic chamber is constructed inside a superconducting shield, filled with liquid helium and maintained at a temperature of 4.2K. An array of magnetic gradiometer probes is immersed entirely in the liquid helium environment of the vacuum cryogenic chamber to achieve a superconducting state. The magnetic gradiometer probe array is specifically composed of three orthogonally placed basic magnetic gradiometer units. Each basic magnetic gradiometer unit contains a pair of spatially fixed superconducting induction coils, which are fixedly separated along the same axis. The spacing between the two superconducting induction coils is preset according to the spatial resolution requirements of the target detection.

[0075] In practical implementation, a pair of superconducting induction coils in each basic magnetic gradiometer unit are driven to operate synchronously to measure the difference in magnetic flux density between two points in space in differential mode. In some embodiments, an independent superconducting quantum interference device (SQU) is configured as the readout circuit for each pair of superconducting induction coils in each basic magnetic gradiometer unit. A superconducting constant current source is used to provide a stable bias operating point for each readout circuit, ensuring that the operating point of the SQU is located in the linear range of the flux-voltage characteristic curve. The original voltage signals output by the readout circuits of each pair of superconducting induction coils are synchronously acquired and input to the positive and negative input terminals of a differential amplifier. The differential amplifier calculates the difference between the two original voltage signals in real time, and this difference, after amplification, directly corresponds to the difference in magnetic flux density between the two spatial points. It can be understood that the relationship between the voltage signal output by the differential amplifier and the difference in magnetic flux density is characterized by the following formula:

[0076]

[0077] in: This represents the difference voltage at the output of the differential amplifier. This represents the output voltage of the readout circuit for the first superconducting induction coil. This represents the output voltage of the readout circuit for the second superconducting induction coil. This represents the gain coefficient of the differential amplifier. The amplified differential voltage signal is then converted from analog to digital, and the resulting digital sequence serves as the effective output of the basic magnetometer unit.

[0078] Optionally, in the example of detecting underground pipelines, when the pipeline is located below a basic magnetic gradiometer unit, the two superconducting induction coils will sense different magnetic induction intensities due to their different distances from the pipeline, resulting in significant difference signals. The raw data sequences output from the three orthogonal basic magnetic gradiometer units are time-aligned and packaged to form a raw magnetic gradient tensor data stream containing three component gradient data. For example, one data frame may contain gradient values ​​in the X, Y, and Z directions, along with their corresponding timestamps. Data alignment is achieved by providing a sampling clock for the analog-to-digital converters of the three basic magnetic gradiometer units using a shared high-precision clock source. This synchronization mechanism ensures that the gradient data in the three directions are strictly corresponding in time.

[0079] In one embodiment of the present invention, see [reference] Figure 2A superconducting transmission line pair is configured to correspond one-to-one with each basic magnetic gradiometer unit in the magnetic gradiometer probe array. The superconducting lead at the output end of each basic magnetic gradiometer unit is connected to the positive and negative input ends of the corresponding superconducting transmission line pair to form a differential signal transmission link. The output end of each pair of superconducting transmission lines is connected to the input end of an independent input channel in the multi-channel SQUID. This input end includes a set of planar spiral input coils for magnetic flux coupling. The mutual inductance coefficient between the planar spiral input coils and the SQUID loop is controlled to be stable within a preset range to ensure effective coupling of the magnetic gradient signal.

[0080] In practical implementation, when the magnetic gradiometer coupled with the SQUID system is used for underground pipeline detection, a superconducting transmission line pair is configured corresponding to each basic magnetic gradiometer unit in the magnetic gradiometer probe array. For example, three basic magnetic gradiometer units corresponding to gradient measurements in the X, Y, and Z directions are each configured with three independent pairs of superconducting transmission lines. The superconducting transmission line pairs are made of niobium-titanium alloy and encapsulated in multiple layers of thermal insulation shielding to maintain the superconducting state in the liquid helium temperature range and suppress external electromagnetic interference. Each pair of superconducting transmission lines includes a forward signal line and a reverse signal line for transmitting differential signals. The superconducting leads at the output end of each basic magnetic gradiometer unit are connected to the forward and reverse input ends of the corresponding superconducting transmission line pairs to form differential signal transmission links. The superconducting leads are led out from the differential amplifier output of the basic magnetic gradiometer unit and electrically connected and mechanically fixed to the ends of the superconducting transmission line pairs through a cryogenic welding process. This connection method ensures that each gradient component signal in the original magnetic gradient tensor data stream is transmitted in differential form.

[0081] In some embodiments, the length of the superconducting transmission line pair is customized according to the detection system layout. For example, the distance from the magnetic gradiometer probe array to the front end of the multichannel SQUID is set to 1.5 meters. At this length, the superconducting transmission line pair must maintain uniform characteristic impedance to minimize signal reflection. The output of each pair of superconducting transmission lines is connected to the input of an independent input channel in the multichannel SQUID. Each independent input channel of the multichannel SQUID contains a set of planar spiral input coils for magnetic flux coupling. These planar spiral input coils are fabricated on a sapphire substrate using a thin-film process and placed near the SQUID chip. In a specific implementation, the output of the superconducting transmission line pair is connected to the access pad of the planar spiral input coil via an indium bonding pad, completing the signal transmission from the transmission line to the input coil. The planar spiral input coil and the SQUID loop achieve signal injection through near-field magnetic flux coupling. The mutual inductance coefficient between the planar spiral input coil and the SQUID loop is stabilized within a preset range. This stability is achieved through a mechanical fixing structure and low-temperature adhesive. The spatial position of the planar spiral input coil relative to the SQUID loop is precisely adjusted and locked with a rigid bracket to prevent displacement during cooling or vibration. It can be understood that the mutual inductance coefficient directly affects the effective coupling efficiency of the magnetic gradient signal. The relationship between the mutual inductance coefficient and the coupling parameters is characterized by the following formula:

[0082]

[0083] in: This represents the coupling coefficient between the planar spiral input coil and the SQUID loop. This represents the mutual inductance coefficient between the planar spiral input coil and the SQUID loop. This represents the self-inductance of the planar spiral input coil. This represents the self-inductance of the SQUID loop. The mutual inductance coefficient is stabilized within a preset range, such as 10 picohens to 50 picohens, to ensure effective coupling of the magnetic gradient signal and avoid overcoupling that could cause signal distortion.

[0084] In some embodiments, for basic magnetometer units with different spatial orientations, the corresponding superconducting transmission line pairs adopt the same specifications to ensure consistent signal transmission characteristics. The DC resistance of the superconducting transmission line pairs is less than 10 nanoohms in the superconducting state to minimize signal attenuation. Optionally, the superconducting transmission line pairs are wrapped with a high-permeability shielding layer to further isolate external alternating magnetic field interference, and the shielding layer is grounded to the system common ground potential. In specific implementations, the input terminal of each independent input channel of the multi-channel SQUID also includes a matching resistor network to optimize the impedance matching between the superconducting transmission line pairs and the planar spiral input coil, reducing the standing wave effect caused by signal reflection. Optionally, after connection, transmission characteristic tests are performed, such as injecting a standard test signal and measuring the signal amplitude and phase after passing through the superconducting transmission line pairs and the planar spiral input coil, comparing the signal difference between direct connection and connection through a transmission link. Data comparison shows that the additional noise introduced by the differential signal transmission link is less than 5% of the system noise floor. This configuration ensures that the raw magnetic gradient tensor data stream is fed from the magnetic gradient meter probe array into the multi-channel SQUID front-end input coil with low loss and low noise, providing a high-fidelity input signal for subsequent signal amplification and processing.

[0085] In one embodiment of the present invention, a preset bias current is applied to the SQUID loop of each SQUID input channel to make it operate in the linear region of the flux-voltage conversion curve. The output voltage signal of the multi-channel SQUID is converted into a feedback current through a feedback resistor. The feedback current drives a tightly coupled feedback coil to generate a compensation flux to counteract the flux change introduced by the input coil, forming a flux-locked loop. The feedback current value of each channel is acquired in real time. The feedback current value is proportional to the input flux change. The feedback current signal of each channel is converted into a voltage signal through a transimpedance amplifier. The amplitude of the voltage signal has been amplified to form a pre-amplified magnetic gradient signal waveform.

[0086] In practical implementation, multi-channel SQUIDs are used in the signal amplification stage of underground pipeline detection systems. A multi-channel SQUID comprises three independent SQUID input channels, each corresponding to a gradient signal input in a spatial direction. A preset bias current is applied to the SQUID loop of each input channel. This preset bias current is provided by a high-precision digital-to-analog converter, and its value is set near the center point of the linear range of the SQUID's flux-voltage conversion curve, ensuring that the output voltage response of the SQUID loop is linearly related to the change in input flux. Applying the preset bias current is a prerequisite for the multi-channel SQUID to enter an effective amplification state.

[0087] In some embodiments, the preset bias current value of the SQUID loop is determined by pre-testing the flux-voltage conversion characteristic curve of each SQUID input channel. During the test, a small alternating flux signal with a known gradient is injected into the SQUID loop, and the output voltage is recorded. The bias current point with the largest output voltage amplitude and no waveform distortion is found as the preset value. The output voltage signal of the multi-channel SQUID is converted into a feedback current through a feedback resistor. The feedback resistor is a low-temperature metal thin-film resistor and is integrated with the SQUID chip on the same substrate to reduce parasitic parameters. The resistance value of the feedback resistor is selected according to the transconductance of the SQUID and the dynamic range required by the system. The feedback current drives a tightly coupled feedback coil. The feedback coil is made with a planar spiral structure and placed close to the SQUID loop to maximize coupling efficiency. The feedback current flowing through the feedback coil generates a compensation flux. The direction of the compensation flux is opposite to the direction of the input flux change introduced by the planar spiral input coil, thereby canceling the input flux change and forming a flux-locked loop. The flux-locked loop forces the operating point of the SQUID loop to always remain at the preset linear point of the flux-voltage conversion curve. Optionally, the stability of the flux-locked loop can be controlled by adjusting the gain parameter of the feedback loop. If the gain parameter is too high, it may cause loop oscillation, while if the gain parameter is too low, it will result in insufficient tracking speed.

[0088] In practical implementation, the feedback current value of each channel is acquired in real time. The acquisition of the feedback current value is achieved through a low-temperature sampling resistor connected in series in the feedback loop. The voltage drop across the sampling resistor is read by a low-temperature preamplifier. The feedback current value is proportional to the change in input magnetic flux. This proportional relationship is determined by the negative feedback mechanism of the magnetic flux lock loop and can be expressed by the following formula:

[0089]

[0090] in: This represents the change in input magnetic flux introduced into the SQUID loop through the planar spiral input coil. This represents the feedback current value collected in real time. This represents a constant proportionality coefficient determined by the mutual inductance coefficient between the feedback coil and the SQUID loop, as well as the total gain of the feedback loop. It can be understood that the above formula shows that after flux locking, the originally weak magnetic flux change is linearly converted into an easily measurable feedback current signal. The feedback current signal of each channel is converted into a voltage signal by a transimpedance amplifier located outside the cryogenic region and having programmable gain. The conversion ratio of the transimpedance amplifier is set according to the expected amplitude range of the feedback current; for example, converting a microamplitude-level feedback current into a volt-level voltage output. The amplitude of the voltage signal has been amplified, forming a pre-amplified magnetic gradient signal waveform.

[0091] In some embodiments, the acquisition and conversion of the feedback current is performed digitally. The voltage signal output from the low-temperature sampling resistor is first digitized by a high-precision analog-to-digital converter. The digitized feedback current data stream is then sent to a digital signal processor (DSP). The DSP internally implements digital transimpedance amplification and outputs a pre-amplified magnetic gradient signal waveform in digital form. Optionally, a low-pass filter is set in the feedback loop to suppress high-frequency noise. The cutoff frequency of the low-pass filter is set to be more than ten times higher than the highest frequency of the magnetic signal that the target pipeline may generate, to ensure that the signal waveform is not distorted. After the above processing, the weak original magnetic gradient tensor data stream coupled from the planar spiral input coil is converted into a voltage waveform signal with significantly enhanced amplitude and a pre-improved signal-to-noise ratio.

[0092] See Figure 3 In this detection scheme, the design of the spiral full-coverage detection route aims to combine the requirements of magnetic gradient hotspot location and safety avoidance. In operation, the detection plane expands outward in a spiral shape from the starting point, covering the pipeline hazard zone containing three magnetic gradient hotspots (corresponding to underground pipelines of different intensities). The route design, including the starting point, inflection point 1, inflection point 2, and the endpoint, ensures complete coverage of the target area by the spiral path while achieving safe detection by avoiding the pink pipeline hazard zone (where excavation is prohibited). The locations of underground pipelines with intensities of 0.8 (purple cross), 0.6 (red cross), and 0.5 (yellow cross) correspond one-to-one with the magnetic gradient hotspot distribution, providing data support for the density and direction of the detection path. During parameter configuration, the spiral path pitch is set to 2m, and the inflection point offset is 1.2 times the grid spacing of the detection plane to balance detection efficiency and hotspot coverage accuracy.

[0093] In one embodiment of the present invention, a reference noise signal waveform generated solely by the environment and the system itself in a targetless state is acquired. A cross-correlation function model is established between the reference noise signal waveform and the noise components in the real-time magnetic gradient signal waveform. Based on this cross-correlation function model, a minimum mean square error algorithm is used to adaptively subtract the estimated noise components from the real-time magnetic gradient signal waveform. For the periodic spike noise patterns unique to the SQUID background noise, a pattern matching and interpolation replacement algorithm is used for filtering. The undenoised magnetic gradient signal is monitored. When isolated noise pulse events are identified, multiple such events are aligned and averaged, and the noise pulse waveform is extracted as a template. In real-time signal processing, the current signal waveform segment is continuously cross-correlated with the noise pulse waveform template. When the cross-correlation value exceeds the dynamic adaptive threshold, the signal waveform segment is determined to be contaminated by periodic spike noise. The start and end time points of the noise pulses in the signal waveform segment are identified. The signal data within the identified contaminated time points are removed, and the normal signal data before and after the time points are reconstructed using spline interpolation algorithm to achieve noise-free replacement. Wavelet transform threshold denoising is performed on the processed signal waveform to further smooth the random noise floor. Finally, a high-fidelity magnetic gradient signal with significantly improved signal-to-noise ratio in both the frequency and time domains is output.

[0094] In practical implementation, for the pre-amplified magnetic gradient signal waveform obtained from underground pipeline detection, a pre-set adaptive digital noise reduction processor performs joint suppression processing of environmental noise and SQUID background noise. A reference noise signal waveform generated solely by the environment and the system itself in a targetless state is acquired. This reference noise signal waveform is obtained by running the magnetic gradiometer-SQUID coupling system at the same detection location, but with all known artificial magnetic sources turned off or far away, and recording its output. The reference noise signal waveform includes geomagnetic noise, industrial frequency interference, and the system's inherent background noise characteristics under specific location conditions. A cross-correlation function model is established between the reference noise signal waveform and the noise components in the real-time magnetic gradient signal waveform. This model quantifies the temporal correlation between the reference noise and the noise components in the real-time signal. Based on the cross-correlation function model, a minimum mean square error algorithm is used to adaptively subtract the estimated noise components from the real-time magnetic gradient signal waveform. The minimum mean square error algorithm iteratively adjusts the weight coefficients of a finite impulse response filter to minimize the cross-correlation energy between the filtered real-time signal and the reference noise, thereby outputting a signal with significantly suppressed noise components.

[0095] In some embodiments, a pattern matching and interpolation replacement algorithm is used to filter out the periodic spike noise patterns unique to the SQUID background noise. The untreated magnetic gradient signal is monitored, and when isolated noise pulse events are identified, multiple such events are aligned and averaged to extract the noise pulse waveform as a template. Template extraction is achieved by aligning and averaging multiple isolated noise pulse events. In real-time signal processing, a sliding cross-correlation calculation is continuously performed between the current signal waveform segment and the noise pulse waveform template. When the cross-correlation value exceeds a dynamic adaptive threshold, the signal waveform segment is determined to be contaminated by periodic spike noise. The dynamic adaptive threshold is dynamically adjusted according to the background noise level. The start and end times of the noise pulses in the signal waveform segment are identified. Signal data within the identified contaminated time points is removed, and the signal for the corresponding time period is reconstructed using normal signal data from the time points before and after the contaminated time points through a spline interpolation algorithm, achieving noise-free replacement. It is understood that the success of pattern matching depends on the accuracy of the template; see Table 1 for details.

[0096] Table 1: Characteristics of Noise Pulse Waveform Templates

[0097]

[0098] Optionally, the matching degree in the sliding cross-correlation calculation can be quantified using the following formula:

[0099]

[0100] in: This indicates the degree of matching between the current signal segment and the noise pulse waveform template. This represents the amplitude value of the current signal segment at the i-th sampling point. This represents the amplitude value of the noise pulse waveform template at the i-th sampling point. This represents the variance estimate of the background noise near the i-th sampling point. Indicates the length of the template. When A match is considered successful when the value is below the dynamic adaptive threshold.

[0101] In some embodiments, the acquisition of the reference noise signal waveform is continuously updated. The adaptive digital noise reduction processor periodically, or when the environmental noise characteristics change significantly, re-acquires a new reference noise signal waveform and updates the cross-correlation function model to maintain adaptive tracking capability against changes in environmental noise. This dynamic update mechanism ensures the robustness of the noise reduction algorithm in changing environments. Optionally, for the removal and interpolation of periodic spike noise, the location and timestamp of each processing event are recorded to form a processing log for subsequent signal quality analysis. After joint suppression processing by the adaptive digital noise reduction processor, the weak magnetic anomaly characteristics related to underground pipelines in the high-fidelity magnetic gradient signal waveform are more clearly revealed.

[0102] See Figure 4 The diagram illustrates the differences in time-domain waveforms between the interpolated signal, the wavelet-denoised signal, and the true valid signal. Specifically, the blue curve represents the interpolated signal, with a signal-to-noise ratio (SNR) of -30.0 dB. Its waveform contains significant random noise and spike interference, with amplitude fluctuations ranging from -5 micro-Φ to 30 micro-Φ, completely masking the subtle variations in the true valid signal. The green curve represents the wavelet-denoised signal, with an SNR improved to -24.4 dB. Through multi-scale decomposition and thresholding using wavelet transform, the random noise floor is significantly smoothed, but due to algorithm characteristics, the overall bias remains around -8 micro-Φ. The orange curve represents the true valid signal, with amplitude fluctuations only between -5 micro-Φ and 5 micro-Φ, reflecting the true characteristics of underground pipeline magnetic anomalies. The comparison shows that after wavelet transform thresholding, high-frequency random noise is effectively suppressed, and the overall waveform fluctuation amplitude is significantly reduced, laying the foundation for subsequent extraction of magnetic anomaly features and target identification.

[0103] In one embodiment of the present invention, high-fidelity magnetic gradient signals from three orthogonal directions are registered according to the spatial geometric position of the magnetic gradient meter probe array. The three-dimensional magnetic field vector distribution at each point in space is calculated using a numerical integration algorithm based on the known mathematical relationship between the magnetic gradient tensor and the magnetic field vector. A spatial point with a known background magnetic field strength is selected as the integration starting point. Based on the measured magnetic gradient tensor data, the change in magnetic field vector between adjacent grid points is calculated step-by-step along a preset spatial integration path. This calculation of the magnetic field vector change is based on the definition of the magnetic gradient tensor, i.e., the rate of change of the magnetic field vector along a certain direction in space is equal to the projection of the magnetic gradient tensor onto that direction. The magnetic field vector change is accumulated along the integration path and combined with the background magnetic field vector at the integration starting point. The magnetic field vectors of each grid point on the integration path are calculated. The calculation process is repeated by changing different integration paths. The calculation results of all paths are fused and corrected by the least squares optimization algorithm to obtain a three-dimensional magnetic field vector distribution with consistent data and minimum error throughout the space. The calculated three-dimensional magnetic field vector distribution is spatially interpolated and gridded. Based on the gridded magnetic field data, all components of the magnetic gradient tensor of each spatial grid point are calculated. The magnitude of the calculated magnetic gradient tensor components or the corresponding components of the maximum principal eigenvalue of the magnetic gradient tensor are mapped to the three-dimensional spatial coordinate system and rendered to generate a three-dimensional contour line or pseudo-color cloud map. The three-dimensional contour line or pseudo-color cloud map is projected onto the two-dimensional observation plane and superimposed with geographical location information to form a high-resolution magnetic gradient distribution map for target identification.

[0104] In practical implementation, the denoised high-fidelity magnetic gradient signal obtained in the underground pipeline detection scenario undergoes gradient reconstruction and spatial resolution enhancement calculations to generate a high-resolution magnetic gradient distribution map for target identification. The high-fidelity magnetic gradient signals in three orthogonal directions are registered according to the spatial geometric position of the magnetic gradiometer probe array. The registration process assigns corresponding spatial position labels to the gradient data collected simultaneously in the three directions based on the precise three-dimensional coordinates of each basic magnetic gradiometer unit within the array. Utilizing the known mathematical relationship between the magnetic gradient tensor and the magnetic field vector, the three-dimensional magnetic field vector distribution at each point in space is calculated using a numerical integration algorithm. A spatial point with a known background magnetic field strength is selected as the integration starting point. This starting point is typically chosen in an area far from the predicted pipeline location, and its background magnetic field vector is pre-measured using a high-precision absolute magnetometer.

[0105] In some embodiments, based on the measured magnetic gradient tensor data, the change in magnetic field vector between adjacent grid points is calculated step by step along a preset spatial integration path. The preset spatial integration path can be a cluster of radial straight lines starting from the integration start point, or it can be a regular grid line covering the detection area. The calculation of the change in magnetic field vector is based on the definition of the magnetic gradient tensor, that is, the rate of change of the magnetic field vector along a certain direction in space is equal to the projection of the magnetic gradient tensor in that direction. This relationship is the core of the numerical integration algorithm and can be expressed by the following formula:

[0106]

[0107] in: Indicates from a spatial point To adjacent spatial points The change in the magnetic field vector, Indicates at point and points The magnetic gradient tensor obtained by interpolation between them Indicates from point Time The displacement vector. By accumulating the changes in the magnetic field vector along the integration path and combining them with the background magnetic field vector at the integration starting point, the magnetic field vector at each grid point on the integration path is calculated.

[0108] In practice, different integration paths are used to repeat the calculation process, and the calculation results of all paths are fused and corrected using a least-squares optimization algorithm. The least-squares optimization algorithm constructs an overdetermined system of equations with the magnetic field vectors of each grid point as unknowns. Each equation in the system corresponds to a magnetic field change constraint on an integration path. Solving this system of equations yields a three-dimensional magnetic field vector distribution that is consistent across the entire space and has the smallest error. The calculated three-dimensional magnetic field vector distribution is then subjected to spatial interpolation and gridding. Spatial interpolation uses the Kriging interpolation algorithm to account for the spatial correlation of the data. Gridding discretizes the continuous space into a regular three-dimensional voxel grid, with each voxel's center point storing a three-dimensional magnetic field vector value. Based on the gridded magnetic field data, all components of the magnetic gradient tensor at each spatial grid point are calculated. The calculation process uses the central difference method to approximate the partial derivatives of the magnetic field vector in each spatial direction.

[0109] It is understood that the magnetic gradient tensor comprises nine spatial derivatives, but five of these components are independent in the passive region. The magnitudes of the calculated magnetic gradient tensor components, or the components corresponding to the largest principal eigenvalues ​​of the magnetic gradient tensor, are mapped onto a three-dimensional coordinate system, generating three-dimensional contour lines or pseudo-color cloud maps. For example, the norm of the magnetic gradient tensor or a component most sensitive to ferrous pipelines can be mapped to color or isosurfaces. The three-dimensional contour lines or pseudo-color cloud maps are then projected onto a two-dimensional observation plane, overlaid with geographic location information, to form a high-resolution magnetic gradient distribution map for target identification. Optionally, the projection process uses vertical projection or perspective projection along a specific viewpoint, with geographic location information derived from GPS or inertial navigation system data acquired synchronously with the magnetic gradient data.

[0110] In some embodiments, the numerical integration algorithm can employ more complex path planning, such as calculating multiple independent paths starting from the integration starting point separately, and then reducing the accumulation of integration errors by weighted averaging at the path intersection points. It is understood that when fusing results from different paths, the least squares optimization algorithm assigns different weights to each path based on its length and the measurement uncertainty of the magnetic gradient tensor data along that path. Optionally, the generated two-dimensional high-resolution magnetic gradient distribution map can be overlaid on a geographic information system base map. In this map, pipeline targets typically appear as linear or chain-like magnetic anomaly stripes along a specific orientation, with significantly improved clarity in outline and magnetic features compared to the original gradient signal map.

[0111] See Figure 5 This image showcases a high-resolution magnetic gradient distribution map generated after gradient reconstruction and spatial resolution enhancement calculations, specifically a pseudo-color cloud map of the magnetic gradient tensor norm at a depth of 6 meters underground. In the image, color variations characterize the spatial distribution intensity of the magnetic gradient tensor norm, and clear linear high-value bands indicate the direction of underground pipelines. This matches the synchronously recorded GPS location information, achieving high-precision positioning and identification of the detected target. Specifically, the image is based on the registered three-dimensional orthogonal magnetic gradient signal. The full-space magnetic field vector distribution is obtained through numerical integration inversion, and then all components of the magnetic gradient tensor are calculated using Kriging interpolation gridding and the central difference method. Finally, the tensor norm is mapped to pseudo-color to reflect its spatial distribution. The color gradient from blue to red in the figure corresponds to the change in the magnetic gradient tensor norm from 0.03 nT / m² to 0.10 nT / m². The red area represents strong magnetic gradient anomalies, which are usually associated with the presence of magnetic targets such as underground ferrous pipelines. The GPS points 1-4 marked in the figure are synchronously acquired geographic reference points used to register the magnetic gradient data with the actual geographic location. The two black dashed lines are the central crosshairs of the detection area, which can help locate the relative position of the anomaly. In addition, the linear magnetic anomaly bands along the pipeline are clearly distinguishable in the figure, indicating that the spatial identification of the magnetic anomaly is significantly improved after gradient reconstruction and resolution enhancement compared to the original signal.

[0112] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for coupling a magnetic gradient meter with a SQUID to improve sensitivity, characterized in that, The following processing steps are included: Deploy at least one set of magnetic gradient meter probe arrays in a superconducting shielded environment to acquire the original magnetic gradient tensor data stream of the target region; The original magnetic gradient tensor data stream is fed into the multi-channel SQUID front-end input coil through a superconducting transmission line. Magnetic flux locking and signal amplification processing are performed in the multi-channel SQUID to generate a magnetic gradient signal waveform that has been preliminarily amplified. The magnetic gradient signal waveform is subjected to joint suppression of environmental noise and SQUID background noise using a pre-set adaptive digital noise reduction processor to generate a noise-reduced high-fidelity magnetic gradient signal. The high-fidelity magnetic gradient signal is subjected to gradient reconstruction and spatial resolution enhancement calculations to generate a high-resolution magnetic gradient distribution map for target recognition.

2. The method for coupling a magnetic gradient meter with a SQUID to improve sensitivity according to claim 1, characterized in that, The step of deploying at least one set of magnetic gradient meter probe arrays in a superconducting shielded environment to acquire the original magnetic gradient tensor data stream of the target region includes: A vacuum cryogenic cavity was constructed inside a superconducting shield, and a set of magnetic gradient meter probe arrays was immersed in the liquid helium environment of the vacuum cryogenic cavity. The magnetic gradiometer probe array consists of three orthogonally placed basic magnetic gradiometer units, each of which contains a pair of spatially fixed superconducting induction coils. The pair of superconducting induction coils in each basic magnetic gradient meter unit are driven to work synchronously to measure the difference in magnetic flux intensity between two points in space in differential mode. Each basic magnetic gradiometer unit outputs in real time the raw data sequence of the difference in magnetic flux density it senses as a function of time; The raw data sequences output by the three orthogonal directions of the basic magnetic gradient meter unit are aligned in time and packaged to form the raw magnetic gradient tensor data stream containing the gradient data of the three components.

3. The method for coupling a magnetic gradient meter with a SQUID to improve sensitivity according to claim 1, characterized in that, The step of feeding the original magnetic gradient tensor data stream into the multi-channel SQUID front-end input coil via a superconducting transmission line includes: Configure a superconducting transmission line pair that corresponds one-to-one with each basic magnetic gradiometer unit in the magnetic gradiometer probe array; The superconducting leads at the output of each basic magnetic gradiometer unit are connected to the positive and negative input ends of the corresponding superconducting transmission line pair to form a differential signal transmission link. The output of each pair of superconducting transmission lines is connected to the input of an independent input channel in a multi-channel SQUID, the input of which includes a set of planar helical input coils for flux coupling. The mutual inductance coefficient between the planar spiral input coil and the SQUID loop is kept stable within a preset range to ensure effective coupling of the magnetic gradient signal.

4. The method for coupling a magnetic gradient meter with a SQUID to improve sensitivity according to claim 1, characterized in that, The process of performing magnetic flux locking and signal amplification in the multi-channel SQUID to generate a pre-amplified magnetic gradient signal waveform includes: A preset bias current is applied to the SQUID loop of each SQUID input channel to make it operate in the linear region of the flux-voltage conversion curve; The output voltage signal of the multi-channel SQUID is converted into a feedback current through a feedback resistor; The feedback current drives a tightly coupled feedback coil to generate a compensating magnetic flux, which counteracts the magnetic flux change introduced by the input coil, thus forming a magnetic flux locking loop. The feedback current value of each channel is collected in real time, and the feedback current value is proportional to the change in input magnetic flux; The feedback current signal of each channel is converted into a voltage signal by a transimpedance amplifier. The amplitude of the voltage signal is amplified to form the pre-amplified magnetic gradient signal waveform.

5. The method for coupling a magnetic gradient meter with a SQUID to improve sensitivity according to claim 1, characterized in that, The process of using a pre-set adaptive digital noise reduction processor to jointly suppress environmental noise and SQUID background noise in the magnetic gradient signal waveform to generate a noise-reduced high-fidelity magnetic gradient signal includes: Acquire reference noise signal waveforms generated solely by the environment and the system itself in a targetless state; Establish a cross-correlation function model between the noise components in the reference noise signal waveform and the real-time magnetic gradient signal waveform; Based on the cross-correlation function model, the estimated noise component is adaptively subtracted from the real-time magnetic gradient signal waveform using the minimum mean square error algorithm. To filter out the periodic spike noise pattern unique to the SQUID background noise, a pattern matching and interpolation replacement algorithm is used. Wavelet transform threshold denoising is performed on the processed signal waveform to further smooth the random noise floor. The final output is a high-fidelity magnetic gradient signal with significantly improved signal-to-noise ratio in both the frequency and time domains.

6. The method for coupling a magnetic gradient meter with a SQUID to improve sensitivity according to claim 1, characterized in that, The step of performing gradient reconstruction and spatial resolution enhancement calculations on the high-fidelity magnetic gradient signal to generate a high-resolution magnetic gradient distribution map for target recognition includes: The high-fidelity magnetic gradient signals in three orthogonal directions are registered according to the spatial geometric position of the magnetic gradient meter probe array; Using the known mathematical relationship between the magnetic gradient tensor and the magnetic field vector, the three-dimensional magnetic field vector distribution at each point in space is calculated by numerical integration algorithm. The calculated three-dimensional magnetic field vector distribution is then subjected to spatial interpolation and meshing. Based on gridded magnetic field data, calculate all components of the magnetic gradient tensor at each spatial grid point; The calculated magnitude of the magnetic gradient tensor components or the component corresponding to the maximum principal eigenvalue of the magnetic gradient tensor is mapped to a three-dimensional spatial coordinate system, and then rendered to generate three-dimensional contour lines or pseudo-color cloud maps. The three-dimensional contour lines or pseudo-color cloud map are projected onto a two-dimensional observation plane, and geographical location information is superimposed to form the high-resolution magnetic gradient distribution map used for target identification.

7. The method for coupling a magnetic gradient meter with a SQUID to improve sensitivity according to claim 2, characterized in that, The method of driving a pair of superconducting induction coils in each basic magnetic gradient meter unit to operate synchronously and measure the difference in magnetic flux density between two points in space in differential mode includes: Each pair of superconducting induction coils in a basic magnetic gradiometer unit is configured with an independent superconducting quantum interference device as a readout circuit; A superconducting constant current source is used to provide a stable bias operating point for each readout circuit; The original voltage signals output by the readout circuits of a pair of superconducting induction coils are acquired synchronously. The original voltage signal is input to the positive and negative input terminals of the differential amplifier; The differential amplifier calculates the difference between the two original voltage signals in real time, and the amplified difference directly corresponds to the difference in magnetic induction intensity between the two spatial points. The amplified difference signal is converted from analog to digital and used as the effective output of the basic magnetic gradient meter unit.

8. The method for coupling a magnetic gradient meter with a SQUID to improve sensitivity according to claim 5, characterized in that, The algorithm employs pattern matching and interpolation replacement to filter out the periodic spike noise patterns unique to the SQUID background noise, including: Monitor the undenoised magnetic gradient signal, and when an isolated noise pulse event is identified, align and average multiple such events to extract the noise pulse waveform as a template. In real-time signal processing, the current signal waveform segment is continuously subjected to sliding cross-correlation calculation with the noise pulse waveform template; When the cross-correlation value exceeds the dynamic adaptive threshold, it is determined that the signal waveform segment is contaminated by the periodic spike noise; Identify the start and end times of noise pulses in the signal waveform segment; The signal data within the identified contaminated time point is removed, and the normal signal data from the time points before and after it is used to reconstruct the signal for the corresponding time period through spline interpolation algorithm, thus achieving noise-free replacement.

9. The method for coupling a magnetic gradient meter with a SQUID to improve sensitivity according to claim 6, characterized in that, The method of utilizing the known mathematical relationship between the magnetic gradient tensor and the magnetic field vector to invert and calculate the three-dimensional magnetic field vector distribution at each point in space through a numerical integration algorithm includes: Choose a spatial point with a known background magnetic field strength as the starting point for integration; Based on the measured magnetic gradient tensor data, the change in magnetic field vector between adjacent grid points is calculated step by step along the preset spatial integration path. The calculation of the change in the magnetic field vector is based on the definition of the magnetic gradient tensor, that is, the rate of change of the magnetic field vector along a certain direction in space is equal to the projection of the magnetic gradient tensor onto that direction. By accumulating the changes in the magnetic field vector along the integration path and combining them with the background magnetic field vector at the integration starting point, the magnetic field vector at each grid point on the integration path is calculated. By repeating the calculation process through different integration paths, and then using a least squares optimization algorithm to fuse and correct the calculation results of all paths, a three-dimensional magnetic field vector distribution with consistent performance and minimal error across the entire space is obtained.

10. A test system for coupling a magnetic gradient meter with a SQUID to improve sensitivity, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method for coupling a magnetic gradient meter with a SQUID to improve sensitivity as described in any one of claims 1 to 9.

Images