A cable line judging method based on multi-information fusion

Abstract

The application relates to a cable line judgment method based on multi-information fusion, which comprises the following steps: collecting ultrasonic signals, electromagnetic signals, current transient signals and impedance response signals of a target line respectively, and performing standardization processing; extracting feature vectors of the signals of various modes, determining a weighting function of the corresponding mode signals based on the extracted feature vectors, performing time domain correlation analysis and frequency domain and phase consistency analysis, respectively obtaining corresponding correlation indexes, and calculating a single mode correlation coefficient based on the correlation indexes of the signals of various modes; weighting and summing the single mode correlation coefficients of the four types of signals to obtain a comprehensive correlation index, wherein the weight of the weighting and summing is adaptively determined; and determining whether the target cable is a same path conductor based on the comprehensive correlation index. Compared with the prior art, the application has the advantages of high line judgment precision, strong signal anti-interference capability and stable judgment process.

Description

Technical Field

[0001] This invention relates to the field of power detection and signal analysis technology, and in particular to a cable identification method based on multi-information fusion. Background Technology

[0002] In fields such as power, communications, and rail transportation, cables serve as critical transmission carriers, existing in vast quantities and laid in complex environments. Traditional cable identification methods often rely on manual experience or single-signal detection, such as using the current response of a single-frequency signal or electromagnetic field induction for identification. However, these methods are susceptible to environmental noise, signal attenuation, and coupling effects between cables, leading to high misidentification rates and insufficient accuracy.

[0003] In recent years, with the development of multi-source signal acquisition and data fusion technologies, cable identification methods based on multi-information fusion have gradually attracted attention. Different physical quantities (such as electromagnetic waves, sound waves, current, voltage, and impedance) can reflect the characteristic attributes and transmission path information of cables from multiple dimensions. For example, ultrasonic signals can reflect mechanical connections and dielectric properties, electromagnetic signals reflect the coupling state of cable conductors, transient current signals reveal continuity characteristics, and impedance response signals characterize the frequency response characteristics of the line. If these multiple signals can be analyzed collaboratively and the correlation between transmitted and received signals can be comprehensively calculated, accurate judgment of cable paths and co-location relationships can be achieved. For example, Chinese patent CN120802134A discloses a method, device, and medium for real-time diagnosis of cable joint resonance. The method includes the following steps: acquiring high-frequency current signals collected by a high-frequency current sensor and ultrasonic vibration signals collected by an ultrasonic sensor array; synchronizing the high-frequency current signals and ultrasonic vibration signals, and extracting high-frequency current features and ultrasonic array features; calculating the main frequency synchronization index, normalized phase stability index, and nonlinear coupling index based on the high-frequency current features and ultrasonic array features, and performing weighted summation to determine the fusion confidence function; determining whether resonance exists based on the fusion confidence function, and if so, locating the resonance source.

[0004] However, existing research still has the following shortcomings: 1. The time synchronization and amplitude matching problems of multi-source signal acquisition have not been effectively solved, resulting in limited fusion accuracy; 2. The lack of robust multimodal correlation analysis algorithms makes it difficult to handle signal recognition in complex noise backgrounds.

[0005] Therefore, there is an urgent need for a method that can simultaneously acquire multiple physical signals and comprehensively determine the cable path through correlation analysis algorithms, so as to improve the reliability and automation level of cable identification. Summary of the Invention

[0006] The purpose of this invention is to provide a cable identification method based on multi-information fusion to solve the problems of low identification accuracy, poor signal anti-interference ability and unstable identification process in the prior art, so as to achieve rapid and accurate identification of target cables.

[0007] The objective of this invention can be achieved through the following technical solutions: According to a first aspect of the present invention, a cable identification method based on multi-information fusion is provided, the method comprising the following steps: The ultrasonic signal, electromagnetic signal, transient current signal, and impedance response signal of the target line were collected and standardized. The feature vectors of each modal signal are extracted, and the weighting function of the corresponding modal signal is determined based on the extracted feature vectors. Time-domain correlation analysis and frequency-domain and phase consistency analysis are performed to obtain the corresponding correlation indexes. The single-mode correlation coefficient is calculated based on the correlation indexes of each modal signal. The single-mode correlation coefficients of the four types of signals are weighted and summed to obtain a comprehensive correlation index, wherein the weights of the weighted summation are adaptively determined. Based on the comprehensive correlation index, it is determined whether the target cables are conductors of the same path.

[0008] The feature vector of the ultrasound signal includes echo delay, envelope peak value, and peak sharpness, wherein the echo delay is calculated as follows: , in, To stimulate the reference signal, To excite the complex conjugate of the reference signal, According to noise power spectral density Whitening frequency domain weights, For regularization parameters, For the frequency domain representation of the echo signal, This represents the inverse Fourier transform. For frequency, For echo delay, For time delay; The method for calculating the peak sharpness is as follows: , in, The sharpness of the envelope peak. The location of the envelope peak. For envelope in Peak value at that location for The second derivative of .

[0009] The feature vector of the electromagnetic signal includes the amplitude spectrum and phase linearity score obtained by performing an FFT transform on the electromagnetic signal. The phase linearity score is calculated as follows: , in, These are the received / transmitted phase spectra, respectively. For time delay, For calculating the frequency bandwidth used for phase consistency, The average phase deviation within the frequency band Ω under the optimal time delay τ. For frequency, The normalized phase linearity consistency score.

[0010] The characteristic vector of the transient current signal includes the wavefront arrival time and the rising edge slope, which are calculated as follows: Highlighting wavefront arrival and upward momentum using the Teager–Kaiser energy operator: , in, The time-domain waveform of the transient current signal. The first derivative of the transient current signal. The second derivative of the transient current signal. For current transient The applied Teager–Kaiser energy operator, For the arrival time of the wavefront, its neighborhood As an indicator of the rising slope.

[0011] The characteristic vector of the impedance response signal includes impedance magnitude, impedance phase, and structural slope. In Bode coordinates, the joint norm of the amplitude-relative slope describes the line structure and termination state. The structural slope is calculated as follows: , in, The magnitude of the impedance. The frequency band for slope calculation. For angular frequency variables, () represents the phase of the impedance. The slope is the structural slope.

[0012] The aforementioned time-domain correlation analysis specifically includes: , in, For modality The time-domain correlation index For modality Weighted cross-correlation value, The time-domain weighting function is determined based on the eigenvectors. , They represent the first k Time-domain waveforms of the transmitter and receiver under each mode. Let be the length of the time-domain integration interval; where, For ultrasound signals, Weighted by envelope amplitude, where, Let be the envelope function of the ultrasound signal. , The peak value is the envelope value. For transient current signals, Weighted by transient energy operators, where, For current transient (t) The Teager–Kaiser energy operator applied This represents the maximum value of the Teager–Kaiser energy operator across the entire time domain; For electromagnetic signals, ,in, Phase difference; For impedance signals ,in, For impedance at frequency The amplitude at that point, This is used to calculate the average impedance amplitude over the frequency band.

[0013] In the aforementioned frequency domain and phase consistency analysis, the cross-spectral density and coherence coefficient are first calculated: , in, and These are the spectrum of the transmitted and received signals, respectively. To receive signals The complex conjugate, For cross-spectral density, The power spectral density of the transmitted signal. The power spectral density of the received signal; Define the band-limited weighted average coherence coefficient : , in, It is a frequency domain weighting function determined based on eigenvectors. For the first k The integral frequency band corresponding to each mode; To measure the degree of matching between phase and time delay, a phase consistency index is defined. : , in, For frequency, For the first k Delay estimates under various modes.

[0014] The weights for the weighted summation are adaptively determined to obtain their optimal values ​​in the following manner: , in, For the first The true label for each sample is 1 for the target cable and 0 for non-target cables. For the sample i modality k The corresponding single-mode correlation coefficient, Modal k The corresponding weights The regularization coefficient is . As the initial empirical weights, This represents the total number of samples.

[0015] The method of determining whether the target cable is a conductor of the same path based on the comprehensive correlation index specifically involves: determining whether the comprehensive correlation index is greater than or equal to the optimal threshold; if so, then it is determined to be a conductor of the same path; otherwise, it is not a conductor of the same path.

[0016] The optimal threshold is obtained by optimization using a joint optimization model based on the dual objectives of minimum mean square error and classification consistency. The objective function of the joint optimization model is defined as: , in, For smooth discriminant function, This is the steepness coefficient. For balancing parameters, and The number of positive and negative samples are respectively. The total number of samples, For the first The true label of each sample For the first The comprehensive correlation index corresponding to each sample The threshold variable to be optimized.

[0017] According to a second aspect of the present invention, a cable identification system based on multi-information fusion is provided, the system comprising: Signal acquisition and preprocessing module: Acquires ultrasonic signals, electromagnetic signals, transient current signals and impedance response signals of the target line respectively, and performs standardization processing; Correlation analysis module: Extracts feature vectors of each modal signal, determines the weighting function of the corresponding modal signal based on the extracted feature vectors, performs time-domain correlation analysis and frequency-domain and phase consistency analysis to obtain the corresponding correlation indexes, calculates the single-mode correlation coefficient based on the correlation indexes of each modal signal, and performs weighted summation of the single-mode correlation coefficients of the four types of signals to obtain the comprehensive correlation index, wherein the weights of the weighted summation are adaptively determined; Output module: Determines whether the target cables are conductors of the same path based on the comprehensive correlation index.

[0018] According to a third aspect of the present invention, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the program to implement the method described thereon.

[0019] According to a fourth aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the method described thereon.

[0020] Compared with the prior art, the present invention has the following beneficial effects: (1) This invention integrates four different physical domain signals: ultrasonic signal, electromagnetic signal, current transient signal and impedance response signal, and extracts feature vector combinations with more physical meaning and line discrimination power for each type of signal. This expands the judgment basis of the same path conductor from a single phenomenon (such as resonance) to multiple essential attributes such as structure, electromagnetic, transient and impedance. It effectively improves the accuracy and anti-interference ability of cable line judgment in complex environments, is not easily affected by a single interference factor, and is suitable for rapid identification of urban power grid, railway communication and industrial control lines. It provides a high-precision and high-reliability technical means for cable detection. (2) The present invention uses a correlation analysis algorithm to achieve multi-dimensional feature fusion, which overcomes the problem of easy misjudgment of single signal line. Furthermore, the introduction of adaptive weights means that the present invention can dynamically adjust the degree of trust in various signals according to factors such as field environment, signal quality, and cable type, making the final comprehensive index more environmentally adaptable and decision-making intelligent. Attached Figure Description

[0021] Figure 1 This is a system structure diagram of the present invention. Figure 2 This is a schematic diagram of the signal acquisition and processing process of the present invention; Figure 3 This is a flowchart of the method of the present invention. Detailed Implementation

[0022] 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, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0023] Unless otherwise defined, the technical or scientific terms used in this application shall have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms “a,” “an,” “an,” “the,” and similar words used in this application do not indicate quantity limitation and may indicate singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units not listed, or may include other steps or units inherent to these processes, methods, products, or devices. The terms “connected,” “linked,” “coupled,” and similar words used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. “Multiple” used in this application refers to two or more. “And / or” describes the relationship between related objects, indicating that three relationships may exist; for example, “A and / or B” can represent: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following objects are in an "or" relationship. The terms "first," "second," and "third" used in this application are merely to distinguish similar objects and do not represent a specific ordering of the objects.

[0024] Example 1 This embodiment provides a cable identification system based on multi-information fusion, such as... Figure 1 As shown, the system includes: Signal acquisition and preprocessing module: Acquires ultrasonic signals, electromagnetic signals, transient current signals and impedance response signals of the target line respectively, and performs standardization processing; Correlation analysis module: Extracts feature vectors of each modal signal, determines the weighting function of the corresponding modal signal based on the extracted feature vectors, performs time-domain correlation analysis and frequency-domain and phase consistency analysis to obtain the corresponding correlation indexes, calculates the single-mode correlation coefficient based on the correlation indexes of each modal signal, and performs weighted summation of the single-mode correlation coefficients of the four types of signals to obtain the comprehensive correlation index, wherein the weights of the weighted summation are adaptively determined; Output module: Determines whether the target cables are conductors of the same path based on the comprehensive correlation index.

[0025] The modules interact with each other via a control bus and data interface, and are uniformly scheduled by a central control unit (MCU) to form a complete cable identification and analysis system. For example... Figure 2 As shown, the system structure adopts a hierarchical modular design, with time synchronization and multimodal fusion as the core, to ensure the consistency of the acquired signals and the reliability of the analysis results.

[0026] During system operation, the signal injection unit in the signal acquisition and preprocessing module generates and injects a preset excitation signal into the target cable. This signal excites the electromagnetic, current, impedance, and acoustic response characteristics of the cable. The signal acquisition unit simultaneously acquires various response signals caused by the excitation signal at the receiving end, including electromagnetic signals, transient current signals, impedance response signals, and ultrasonic signals. The signal preprocessing unit filters, denoises, normalizes amplitude and phase, and aligns the acquired raw data in the time domain to obtain a signal input with a uniform scale. The correlation analysis module establishes a multimodal information fusion criterion by calculating the time domain and frequency domain correlation indexes between different signals. The output module performs intelligent discrimination based on the fusion results and the judgment threshold, and displays the judgment conclusion and confidence result on the host computer interface in real time.

[0027] The system uses a LoRa communication link as its synchronization core. The LoRa module is responsible for transmitting time synchronization frames and status information between the transmitter and receiver, enabling long-distance wireless synchronization sampling. Its communication bandwidth is 125kHz, the transmission distance can reach 1km, and the synchronization accuracy is better than 1 microsecond. The system adopts a two-layer structure design: the bottom layer is the hardware signal link, which completes data injection and acquisition; the upper layer is the algorithm fusion and decision logic, realizing comprehensive judgment based on multiple information. Through this architecture, this invention can uniformly process multi-physical field signals in complex electromagnetic environments, significantly improving the accuracy and stability of cable path identification.

[0028] The signal injection unit is the core unit of the signal source in this invention, mainly comprising a signal source control unit, a power amplification unit, a coupling injection unit, and a protection circuit. The signal source control unit uses a microcontroller to generate various excitation waveforms, including sweep signals, step signals, and pulse signals, to stimulate the cable's response characteristics in different physical domains. The power amplification unit boosts the low-amplitude signal to an amplitude range suitable for driving the cable (typical output power 2–5W), and connects it to the cable port via an isolation transformer or capacitive coupling to avoid ground loops and signal reflections.

[0029] In practical applications, the excitation signals are divided into the following four categories: (1) Electromagnetic excitation signal (C1): used to excite the electromagnetic response between the outer conductor of the cable and the surrounding medium; (2) Transient current excitation signal (C2): used to detect the effect of sudden changes in current in the conductor on the joint and insulation structure; (3) Impedance response excitation signal (C3): The complex impedance characteristics of the cable are obtained by sweep frequency voltage excitation; (4) Ultrasonic excitation signal (C4): used to excite the acoustic propagation characteristics inside the cable and the insulation layer.

[0030] The signal source control unit uses a high-precision digital-to-analog converter (DAC) to output the excitation waveform, and its time base is kept consistent with the LoRa synchronization clock to ensure time alignment between signal injection and acquisition. Overvoltage protection and isolation circuits are installed at the injection end to prevent reverse coupling interference. To ensure the spectral purity of the signal, the signal is preprocessed by a bandpass filter before output, with a bandwidth set to 4–60 kHz to avoid power system frequency interference and high-frequency noise.

[0031] During signal injection, the system first issues a trigger command from the central control unit, and the signal source generates a set waveform signal and injects it into the cable. After injection, the LoRa module immediately sends a synchronization trigger frame to the receiver, enabling the sampling device to start data acquisition within the same time window. This design ensures the timing consistency of signal injection and acquisition, eliminating phase drift caused by transmission delay. Through the joint excitation of signals from multiple physical domains, the system can simultaneously capture the electromagnetic field, acoustic field, and impedance response information of the cable in a single injection, thereby significantly improving the signal integrity and feature discrimination of the cable identification algorithm.

[0032] The advantages of this module are: first, it avoids the information loss problem caused by a single excitation source in traditional line detection methods; second, it improves the time-frequency consistency of the line detection signal through parallel injection of multiple signals and high-precision synchronous control; and third, the system has a simple structure, low power consumption, and can adapt to the detection needs of complex field environments and different cable types.

[0033] The signal acquisition unit is used to acquire the multimodal response signals of the cable port at synchronous moments and convert them into digital signals for input to the subsequent signal preprocessing unit. The signal acquisition unit includes an electromagnetic induction receiving unit, a current sensing unit, an impedance measurement unit, an ultrasonic receiving unit, an analog-to-digital conversion circuit, and a synchronization control unit. Each receiving unit is connected to the A / D conversion module through an isolation amplifier circuit to achieve multi-channel synchronous sampling.

[0034] In this invention, the electromagnetic induction receiving unit employs a high-sensitivity coil sensor to detect changes in the external magnetic field of the cable; the current sensing unit uses a Hall current probe or a resistance sampling structure to capture transient current signals within the conductor; the impedance measurement unit consists of a swept frequency source and a measurement bridge, obtaining complex impedance parameters by measuring the voltage-to-current ratio; and the ultrasonic receiving unit consists of a piezoelectric sensor array to receive acoustic signals propagating inside the cable. The signals acquired by each unit are differentially amplified and low-pass filtered before being input to a multi-channel sampling card, with a sampling accuracy of 16 bits and a maximum sampling rate of 1 MHz.

[0035] To achieve high-precision synchronous sampling, the system utilizes LoRa communication for time alignment. Delay correction is performed between the transmitter and receiver using a timestamp synchronization mechanism. The propagation delay calculation formula is as follows: , in, The timing of the transmission of the synchronization frame at the transmitting end. This refers to the receiving time at the receiving end. The system uses... As a compensation for bidirectional propagation delay, the sampling trigger clock is corrected to ensure that the time difference between the two sampling ends is less than 1 microsecond. To reduce the influence of environmental factors, a temperature and humidity compensation model is introduced into the system, and the propagation speed is corrected using the dielectric constant. ,in , At the speed of light, This is an empirical coefficient. For temperature, Humidity.

[0036] Data from each channel in the signal acquisition unit is transmitted to the central control unit in real time via the SPI bus. The system supports multi-threaded data buffering and parallel processing. The acquired data includes not only the raw waveform but also sampling timestamps, channel numbers, and synchronization status indicators for subsequent phase calibration and signal alignment. Through this mechanism, the system can stably obtain high-fidelity signals in complex electromagnetic interference environments, ensuring that data from different physical channels can be compared under the same time reference frame.

[0037] The technical advantages of this module are: it achieves low-cost, high-precision parallel acquisition of multiple signals through a wireless synchronous acquisition scheme; it avoids signal line interference and grounding differences in traditional wired synchronous methods; and it can uniformly acquire multiple physical signals over a wide frequency band, providing a stable and accurate data source for subsequent signal processing and feature fusion.

[0038] The signal preprocessing unit mainly includes a filtering unit, a normalization unit, a phase calibration unit, and a buffer control unit. This module's function is to standardize the raw signal output from the acquisition module to eliminate amplitude differences, timing shifts, and noise interference between different signal sources and sensing channels, enabling all signals to participate in fusion analysis under a unified time reference and amplitude / phase scale.

[0039] The signal preprocessing unit operates as follows: First, the filtering unit performs noise suppression and frequency band constraint on the input signal. The system uses a bandpass filter to limit the effective bandwidth to 4–60 kHz to remove low-frequency power frequency components and high-frequency environmental interference; 50 Hz and 150 Hz notch filters are used to further improve signal purity against power harmonic interference. Filter parameters are automatically calculated based on the sampling frequency to meet the Nyquist criterion.

[0040] Secondly, the normalization unit performs amplitude and energy normalization on the filtered signal. Let the first... Channel signal is Its mean is The standard deviation is Then the normalization formula is: , Through this operation, each signal channel is converted to a uniform scale with zero mean and unit variance, thereby eliminating the influence of sensor sensitivity and gain errors.

[0041] Furthermore, the system utilizes a phase calibration unit to achieve phase consistency of multimodal signals. To eliminate sensor link delay and phase drift, a Hilbert transform is employed to obtain the analytic signal, and time-domain alignment is achieved through instantaneous phase calculation. , in, This is the Hilbert operator. This method can achieve synchronous phase correction in the time and frequency domains, enabling comparison of electromagnetic, ultrasonic, current, and impedance signals in the same phase reference frame.

[0042] Finally, the cache control unit is used to control the stored signal information.

[0043] The correlation analysis module is the core functional unit of this invention for multi-signal fusion and cable identification. It mainly includes a feature extraction unit, a single-mode correlation calculation unit, a multi-mode fusion unit, a threshold calculation unit, and a buffered output unit. The task of this module is to establish a feature-based correlation model between signals from different physical domains, and to identify and distinguish the target cable by quantifying the degree of matching.

[0044] Based on the features output by the feature extraction unit (e.g., envelope peak value, echo delay, phase consistency), a weighting mechanism is introduced for each signal mode (ultrasonic signal, transient current signal, electromagnetic signal, impedance response signal). The features of each signal serve as a weighting function, adjusting the contribution of different signal segments to the correlation calculation. Thus, when calculating correlation, the previously extracted features directly affect the calculation results, ensuring an organic link between feature extraction and subsequent analysis.

[0045] The specific implementation process of the correlation analysis module is described in detail in Example 2.

[0046] The output module is the decision-making execution part of this invention, mainly composed of a logic line judgment unit, a confidence calculation unit, a result output unit, and a communication interface unit. Based on the comprehensive indicators and thresholds output by the correlation analysis module, this module automatically determines whether the cable paths are consistent and outputs the results.

[0047] First, the logic line detection unit executes the core discrimination rule: when the comprehensive correlation index is greater than or equal to the optimal threshold, the system determines that the cable under test is the target cable; when the comprehensive correlation index is less than the optimal threshold, it is determined to be a non-target cable. To improve robustness, the system adopts a sliding window averaging strategy, performing a weighted average of multiple consecutive sampling results. Only when the results are consistent across multiple samplings is the final line detection conclusion output, thereby effectively suppressing misjudgments caused by occasional noise interference.

[0048] Secondly, the confidence calculation unit is used to measure the reliability of the line judgment result, and the calculation formula is as follows: , in, and These represent the minimum and maximum values ​​of the comprehensive correlation index in the historical dataset, respectively. Confidence level. The value range is [0,1], and a larger value indicates a more reliable matching result. When the system marks it as a "high confidence threshold"; when When the test is completed, the system prompts the user to perform a retest. This method improves the interpretability of the results while maintaining the speed of line determination.

[0049] Secondly, the result output unit is responsible for packaging the cable identification results, comprehensive correlation, threshold, and confidence information into a unified package, and uploading it to the host computer or remote monitoring terminal via the communication interface. The host computer can display the identification results of "target cable" or "non-target cable" in real time, and simultaneously display the correlation curve and confidence value in a graphical interface, making it easy for maintenance personnel to intuitively judge the consistency of the cable path. The communication interface supports both LoRa wireless transmission and wired serial transmission to meet the needs of different application scenarios.

[0050] The technical advantage of this module lies in its integrated line determination mechanism, which combines automatic line determination, confidence assessment, and result visualization. It can automatically complete judgment and recording under unattended conditions. Compared to traditional manual line determination, this module significantly improves detection efficiency and result reliability, achieving a line determination accuracy rate of over 97%, and possesses high real-time performance and intelligent features.

[0051] The system's workflow is as follows: (1) Signal injection: The system injects a preset excitation signal into the target cable through the signal injection unit. The signal types include frequency sweep, step and pulse signals, which are used to excite the cable’s response in the electromagnetic, impedance, current and acoustic domains.

[0052] (2) Synchronous acquisition: The LoRa module is responsible for synchronous control, so that the acquisition end and the injection end start data acquisition at the same time. The signal acquisition unit simultaneously acquires four types of response signals, namely electromagnetic signals, transient current signals, impedance response signals and ultrasonic signals, and timestamps them.

[0053] (3) Signal processing: The acquired raw signals are filtered, normalized and phase corrected to unify the dimensions and phase. The bandpass filter suppresses low-frequency interference and the normalization algorithm eliminates amplitude differences to ensure that different signals are comparable.

[0054] (4) Correlation analysis: The system calculates the cross-correlation coefficient between the transmitted and received signals and obtains the comprehensive correlation index through a weighted fusion algorithm.

[0055] (5) Cable Judgment and Output: The output module makes a judgment based on the comprehensive correlation index. It outputs a conclusion of "target cable" or "non-target cable" and calculates the confidence level. This serves as a reliability metric. The results are transmitted wirelessly in real time to a host computer interface for display, completing the entire testing loop.

[0056] Through the above process, the system automates signal excitation, synchronous acquisition, feature extraction, fusion analysis, and intelligent line determination. The entire line determination process takes 1-2 seconds, enabling rapid and accurate on-site detection.

[0057] Compared with existing technologies, this system has significant advantages in signal synchronization, information fusion, algorithm intelligence, and system reliability: 1. High-precision synchronization performance: With LoRa wireless synchronization sampling technology, the clock deviation between the transmitter and receiver is less than 1 microsecond, which completely solves the time error problem caused by signal line length and grounding differences in traditional wired synchronization.

[0058] 2. Strong multi-information fusion and recognition capabilities: By employing a multimodal information fusion technology that integrates electromagnetic signals, transient current signals, impedance response signals, and ultrasonic signals, and utilizing a weighted correlation model to fully exploit the complementarity of features from different physical domains, the system maintains stable line detection capability even in high-noise environments.

[0059] 3. The algorithm has strong self-learning and environmental adaptability: The fusion algorithm can dynamically adjust the weights and thresholds according to the current signal-to-noise ratio, has adaptive learning capabilities, and can maintain stable performance when the cable type changes or external interference conditions change.

[0060] 4. The system has a simple structure and low power consumption: The modular design combined with LoRa communication significantly reduces hardware complexity and power consumption. The overall system power consumption is less than 1W, and it can be powered by a portable power supply for extended periods, making it suitable for use in the field and confined spaces.

[0061] 5. Enhanced visualization and remote monitoring capabilities: The host computer displays the correlation curve, line judgment results, and confidence value in real time, supports remote monitoring and data archiving, and facilitates cable maintenance personnel to quickly locate and judge the line.

[0062] Example 2 This embodiment provides a cable identification method based on multi-information fusion, building upon Embodiment 1. Figure 3 As shown, the method includes the following steps: S1 collects the ultrasonic signal, electromagnetic signal, transient current signal and impedance response signal of the target line respectively, and performs standardization processing.

[0063] S2, extract the feature vectors of each modal signal, determine the weighting function of the corresponding modal signal based on the extracted feature vectors, perform time-domain correlation analysis and frequency-domain and phase consistency analysis, obtain the corresponding correlation indexes, and calculate the single-mode correlation coefficient based on the correlation indexes of each modal signal.

[0064] S21 extracts a small number of core indicators with strong discriminative power from the normalized signal by mode and feature domain, and forms a unified feature vector for subsequent correlation analysis. Let the preprocessed signal be... Channel signal is Its analytical form is The envelope is Frequency domain representation is .

[0065] 1) Ultrasound signal (US) - time-domain dominant, supplemented by time-frequency robustness.

[0066] To simultaneously obtain echo delay and envelope peak characteristics, a weighted matching "whitening cross-correlation" and peak sharpness are used to jointly describe the characteristics of the ultrasound signal, resulting in an eigenvector that includes echo delay, envelope peak value, and peak sharpness.

[0067] The echo delay is calculated as follows: , in, To stimulate the reference signal, To excite the complex conjugate of the reference signal, According to noise power spectral density Whitening frequency domain weights, For regularization parameters, For the frequency domain representation of the echo signal, This represents the inverse Fourier transform. For frequency, For echo delay, For time delay.

[0068] The method for calculating peak sharpness is as follows: , in, The sharpness (curvature) of the envelope peak. The location of the envelope peak. For envelope in Peak value at that location for The second derivative of . The sharper the peak ( The larger the value, the clearer the echo from the same cable.

[0069] 2) Electromagnetic signals (EM) – primarily based on frequency domain phase consistency.

[0070] The eigenvectors of an electromagnetic signal include the amplitude spectrum and phase linearity score obtained by performing an FFT transform on the electromagnetic signal.

[0071] Specifically, based on the frequency domain representation of the received EM signal The frequency domain feature vector of this channel It consists of two parts: one is composed of The sampled amplitude spectrum features are used to characterize the energy distribution at each frequency point; secondly, the calculated phase linearity consistency index is used... Characterized by a single scalar In frequency band The magnitude of the residual. Ultimately, it can be expressed as: , in The amplitude spectrum obtained from the FFT of the received electromagnetic signal, and the aforementioned phase spectrum They belong to the same frequency domain representation.

[0072] To suppress the effects of coupling and dispersion, phase linearity consistency is used as the core indicator: , in, These are the received / transmitted phase spectra, respectively. For time delay, For calculating the frequency bandwidth used for phase consistency, frequency band The bandwidth (i.e., the length of the integration interval); The average phase deviation within the frequency band Ω under the optimal time delay τ reflects the degree of residual mismatch between the received phase and the reference phase. The smaller the value, the closer the phase is to linearity and the better the consistency. For frequency, For linear phase terms, corresponding to a phase model with an ideal time delay of τ; The normalized phase linearity consistency score is denoted as 1. The closer the value is to 1, the more linear the phase and the clearer the features.

[0073] 3) Current transient signal (IT) - edge dynamics.

[0074] The eigenvectors of a transient current signal include the wavefront arrival time and the rising edge slope, which are calculated as follows: Highlighting wavefront arrival and upward momentum using the Teager–Kaiser energy operator: , in, The time-domain waveform of the transient current signal. The first derivative of the transient current signal. The second derivative of the transient current signal. For current transient The applied Teager–Kaiser energy operator, For the time point corresponding to the Teager–Kaiser energy peak, i.e., the time-of-arrival arrival time, its neighborhood... As an indicator of the rising slope.

[0075] 4) Impedance response signal (Z) – the structural slope of the amplitude-phase combination.

[0076] The eigenvectors of the impedance response signal include impedance magnitude, impedance phase, and structural slope. In Bode coordinates, the joint norm of the magnitude-relative slope describes the line structure and termination state. The structural slope is calculated as follows: , in, The magnitude of the impedance. The frequency band for slope calculation. For angular frequency variables, () represents the phase of the impedance. It measures the overall curvature / slope intensity of both amplitude and phase frequencies, and can distinguish between cables with the same slope (smooth and consistent) and cables with different slopes / abnormal terminations (abrupt slope). This represents the logarithmic slope of the amplitude-frequency response in Bode coordinates. This represents the logarithmic slope of the phase frequency response in Bode coordinates.

[0077] Each modality retains only a small number of core metrics to form a feature vector, which is then cached according to timestamps. The final assembly is as follows This choice significantly improves the discriminative power under conditions of strong noise, coupling, and varying termination, while maintaining controllable feature dimensions and meeting real-time requirements.

[0078] S22, Time-domain correlation analysis.

[0079] The time-domain correlation analysis specifically includes: , in, For modality The time-domain correlation index For modality Weighted cross-correlation value, The time-domain weighting function is determined based on the eigenvectors. , They represent the first k Time-domain waveforms of the transmitter and receiver under each mode. The length of the time-domain integration interval; the time-domain correlation index. Indicates the first The temporal matching degree of each modality ensures that the contribution of features is properly reflected in the correlation calculation.

[0080] Among them, for ultrasound signals, Weighted by envelope amplitude, where, Let be the envelope function of the ultrasound signal. , This is the maximum amplitude (peak value) of the envelope within the observation interval. For transient current signals, Weighted by transient energy operators, where, For current transient (t) The Teager–Kaiser energy operator applied This represents the maximum value of the Teager–Kaiser energy operator across the entire time domain; For electromagnetic signals, ,in, Phase difference; For impedance signals ,in, For impedance at frequency The amplitude at that point, This is used to calculate the average impedance amplitude over the frequency band.

[0081] S23, Frequency Domain and Phase Consistency Analysis.

[0082] In this step, to balance frequency domain characteristics and phase response consistency, the cross-spectral density and coherence coefficient are first calculated: , in, and These are the spectrum of the transmitted and received signals, respectively. To receive signals The complex conjugate, For cross-spectral density, The power spectral density of the transmitted signal. This represents the power spectral density of the received signal.

[0083] Define the band-limited weighted average coherence coefficient : , in, It is a frequency domain weighting function determined based on eigenvectors. For the first k The integral frequency band corresponding to each mode.

[0084] In this embodiment, The specific determination method is as follows: First, select the effective frequency band using the various frequency domain / impedance domain characteristics obtained in the preceding steps: for example, for electromagnetic signals, retain the phase linearity consistency index. Greater than the threshold The phase-stable frequency band; for impedance signals, retain the impedance amplitude and phase slope index. Less than the threshold The smooth frequency bands. Then within these frequency bands, let... Take 1 (or normalize the weights within the range of 0 to 1 according to the eigenvalue), and set the weights in the remaining frequency bands as follows: This highlights the importance of high-quality frequency bands. The contribution. Then, for The integral frequency band corresponding to non-zero values, i.e., the first... The effective frequency band of each mode.

[0085] In addition, a phase consistency index is defined to measure the degree of matching between phase and time delay. : , in, For frequency, For the first k Time delay estimates under various modes. When the phase difference can be passed through the linear time delay term... When fitting precisely, A value close to 1 indicates that the signal is coherent and consistent in this mode.

[0086] S24, calculate the single-mode correlation coefficient.

[0087] To standardize the measurement, three metrics (time-domain correlation) are used for different modalities. Frequency domain coherence Phase consistency The correlation coefficients are then weighted and fused into single-mode correlation coefficients. , in , represents the weight, and The weights can be dynamically adjusted based on the signal-to-noise ratio, peak stability, and characteristic confidence level. This index comprehensively considers the amplitude, frequency, and phase characteristics of each modal signal and is the basis for line determination calculation.

[0088] S3, the single-mode correlation coefficients of the four types of signals are weighted and summed to obtain a comprehensive correlation index, wherein the weights of the weighted summation are adaptively determined.

[0089] The single-mode correlation coefficients of the four types of signals The comprehensive relevance index is obtained by weighting and summing the results according to their importance: , in, Modal k The corresponding weights.

[0090] The weights for the weighted summation are adaptively determined using the following method: , in, For the first The true label for each sample is 1 for the target cable and 0 for non-target cables. For the sample i modality k The corresponding single-mode correlation coefficient, Modalk The corresponding weights The regularization coefficient is . As the initial empirical weights, The total number of samples. This optimization model simultaneously constrains weights and sample error during iterative training, ensuring that high-confidence modes receive larger weights and low-signal-noise modes automatically decay. The final optimal weight vector is obtained. These are the parameters of the multimodal fusion model.

[0091] S4. Based on the comprehensive correlation index, determine whether the target cable is a conductor of the same path.

[0092] Specifically, the process involves determining whether the comprehensive correlation index is greater than or equal to the optimal threshold. If so, the conductors are considered to be on the same path; otherwise, they are not. To improve robustness, a sliding window averaging strategy is employed, weighted averaging of multiple consecutive sampling results. Only when the results from multiple samplings are consistent is the final line judgment output, thereby effectively suppressing misjudgments caused by occasional noise interference.

[0093] The optimal threshold is obtained by optimization using a joint optimization model based on the dual objectives of minimum mean square error and classification consistency. The objective function of the joint optimization model is defined as: , in, For smooth discriminant function, This is the steepness coefficient. For balancing parameters, and The number of positive and negative samples are respectively. The total number of samples, For the first The true label of each sample For the first The comprehensive correlation index corresponding to each sample The threshold variable to be optimized.

[0094] This model simultaneously minimizes the prediction error and the overlap of the class distribution, obtained through gradient descent or quasi-Newton iteration. .

[0095] Note: Unlike the traditional minimum variance threshold, this optimization model explicitly introduces a smoothing discriminant function. By separating the inter-class terms, a self-learning threshold model is formed during training. This means that weight optimization and threshold training are two independent subsystems: the former learns the importance of each modality, while the latter learns the optimal decision boundary on the fused output. The two are coupled together through a comprehensive index.

[0096] Experimental results show that this method can still maintain a line detection accuracy of over 95% in complex environments with a signal-to-noise ratio as low as 5dB, significantly improving the system's anti-interference capability and environmental adaptability, and providing a highly reliable algorithmic basis for intelligent cable line detection.

[0097] Example 3 The electronic device of this invention includes a central processing unit (CPU), which can perform various appropriate actions and processes according to computer program instructions stored in read-only memory (ROM) or loaded from a storage unit into random access memory (RAM). The RAM may also store various programs and data required for device operation. The CPU, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0098] Multiple components in the device are connected to the I / O interface, including: input units such as keyboards and mice; output units such as various types of displays and speakers; storage units such as disks and optical discs; and communication units such as network interface cards (NICs), modems, and wireless transceivers. The communication unit allows the device to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0099] The processing unit executes the various methods and processes described above, such as methods S1 to S4. For example, in some embodiments, methods S1 to S4 may be implemented as computer software programs tangibly contained in a machine-readable medium, such as a storage unit. In some embodiments, part or all of the computer program may be loaded and / or installed on the device via ROM and / or a communication unit. When the computer program is loaded into RAM and executed by the CPU, one or more steps of methods S1 to S4 described above may be performed. Alternatively, in other embodiments, the CPU may be configured to execute methods S1 to S4 by any other suitable means (e.g., by means of firmware).

[0100] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0101] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0102] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0103] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A cable identification method based on multi-information fusion, characterized in that, The method includes the following steps: The ultrasonic signal, electromagnetic signal, transient current signal, and impedance response signal of the target line were collected and standardized. The feature vectors of each modal signal are extracted, and the weighting function of the corresponding modal signal is determined based on the extracted feature vectors. Time-domain correlation analysis and frequency-domain and phase consistency analysis are performed to obtain the corresponding correlation indexes. The single-mode correlation coefficient is calculated based on the correlation indexes of each modal signal. The single-mode correlation coefficients of the four types of signals are weighted and summed to obtain a comprehensive correlation index, wherein the weights of the weighted summation are adaptively determined. Based on the comprehensive correlation index, it is determined whether the target cables are conductors of the same path.

2. The cable identification method based on multi-information fusion according to claim 1, characterized in that, The feature vector of the ultrasound signal includes echo delay, envelope peak value, and peak sharpness, wherein the echo delay is calculated as follows: ， in, To stimulate the reference signal, To excite the complex conjugate of the reference signal, According to noise power spectral density Whitening frequency domain weights, For regularization parameters, For the frequency domain representation of the echo signal, This represents the inverse Fourier transform. For frequency, For echo delay, For time delay; The method for calculating the peak sharpness is as follows: ， in, The sharpness of the envelope peak. The location of the envelope peak. For envelope in Peak value at that location for The second derivative of .

3. The cable identification method based on multi-information fusion according to claim 1, characterized in that, The feature vector of the electromagnetic signal includes the amplitude spectrum and phase linearity score obtained by performing an FFT transform on the electromagnetic signal. The phase linearity score is calculated as follows: ， in, These are the received / transmitted phase spectra, respectively. For time delay, For calculating the frequency bandwidth used for phase consistency, The average phase deviation within the frequency band Ω under the optimal time delay τ. For frequency, The normalized phase linearity consistency score.

4. The cable identification method based on multi-information fusion according to claim 1, characterized in that, The characteristic vector of the transient current signal includes the wavefront arrival time and the rising edge slope, which are calculated as follows: Highlighting wavefront arrival and upward momentum using the Teager–Kaiser energy operator: ， in, The time-domain waveform of the transient current signal. The first derivative of the transient current signal. The second derivative of the transient current signal. For current transient The applied Teager–Kaiser energy operator, For the arrival time of the wavefront, its neighborhood As an indicator of the rising slope.

5. The cable identification method based on multi-information fusion according to claim 1, characterized in that, The characteristic vector of the impedance response signal includes impedance magnitude, impedance phase, and structural slope. In Bode coordinates, the joint norm of the amplitude-relative slope describes the line structure and termination state. The structural slope is calculated as follows: ， in, The magnitude of the impedance. The frequency band for slope calculation. For angular frequency variables, () represents the phase of the impedance. The slope is the structural slope.

6. The cable identification method based on multi-information fusion according to claim 1, characterized in that, The aforementioned time-domain correlation analysis specifically includes: ， in, For modality The time-domain correlation index For modality Weighted cross-correlation value, The time-domain weighting function is determined based on the eigenvectors. , They represent the first k Time-domain waveforms of the transmitter and receiver under each mode. Let be the length of the time-domain integration interval; where, For ultrasound signals, Weighted by envelope amplitude, where, Let be the envelope function of the ultrasound signal. , The peak value is the envelope value. For transient current signals, Weighted by transient energy operators, where, For current transient (t) The Teager–Kaiser energy operator applied This represents the maximum value of the Teager–Kaiser energy operator across the entire time domain; For electromagnetic signals, ,in, Phase difference; For impedance signals ,in, For impedance at frequency The amplitude at that point, This is used to calculate the average impedance amplitude over the frequency band.

7. The cable identification method based on multi-information fusion according to claim 1, characterized in that, In the aforementioned frequency domain and phase consistency analysis, the cross-spectral density and coherence coefficient are first calculated: ， in, and These are the spectrum of the transmitted and received signals, respectively. To receive signals The complex conjugate, For cross-spectral density, The power spectral density of the transmitted signal. The power spectral density of the received signal; Define the band-limited weighted average coherence coefficient : ， in, It is a frequency domain weighting function determined based on eigenvectors. For the first k The integral frequency band corresponding to each mode; To measure the degree of matching between phase and time delay, a phase consistency index is defined. : ， in, For frequency, For the first k Delay estimates under various modes.

8. The cable identification method based on multi-information fusion according to claim 1, characterized in that, The weights for the weighted summation are adaptively determined to obtain their optimal values ​​in the following manner: ， in, For the first The true label for each sample is 1 for the target cable and 0 for non-target cables. For the sample i modality k The corresponding single-mode correlation coefficient, Modal k The corresponding weights The regularization coefficient is . As the initial empirical weights, This represents the total number of samples.

9. The cable identification method based on multi-information fusion according to claim 1, characterized in that, The method of determining whether the target cable is a conductor of the same path based on the comprehensive correlation index specifically involves: determining whether the comprehensive correlation index is greater than or equal to the optimal threshold; if so, then it is determined to be a conductor of the same path; otherwise, it is not a conductor of the same path.

10. A cable identification method based on multi-information fusion according to claim 9, characterized in that, The optimal threshold is obtained by optimization using a joint optimization model based on the dual objectives of minimum mean square error and classification consistency. The objective function of the joint optimization model is defined as: ， in, For smooth discriminant function, This is the steepness coefficient. For balancing parameters, and The number of positive and negative samples are respectively. The total number of samples, For the first The true label of each sample For the first The comprehensive correlation index corresponding to each sample The threshold variable to be optimized.

Images