Method for dynamically determining perimeter alarm threshold of detection cable based on double-cable coupling

Abstract

The application discloses a double-cable coupling-based detection cable perimeter alarm threshold dynamic determination method and relates to the technical field of detection cable perimeter alarm, and comprises the following steps: on the basis of a spectrum jump criterion set, the consistency curvature of a main frequency dispersion point track of a double-cable coupling signal is calculated, the periodic repetition density and the energy return ratio are analyzed, and it is judged whether the detection cable passes through an electromagnetic reflection structure; in the case that the detection cable passes through the electromagnetic reflection structure, the relative slip distribution of signal channels corresponding to the first cable and the second cable in the double-cable coupling signal on the frequency axis is extracted, and a feature vector set used for describing the spectrum misregistration characteristics of the double-cable coupling signal is generated. The application solves the problem that the alarm determination is inaccurate due to spectrum misregistration when the double-cable coupling detection cable passes through the electromagnetic reflection structure, and realizes accurate determination and self-adaptive dynamic regulation and control of the perimeter alarm threshold in a complex structure environment.

Description

Dynamic Determination Method for Perimeter Alarm Threshold of Detection Cable Based on Dual-Cable Coupling Technical Field

[0001] This invention relates to the field of perimeter alarm technology for detection cables, and specifically to a method for dynamically determining the perimeter alarm threshold of detection cables based on dual-cable coupling. Background Technology

[0002] Dynamic threshold determination of perimeter alarm based on dual-cable coupling refers to the use of signal coupling between two detection cables deployed on the same physical boundary in a perimeter intrusion detection system to achieve highly sensitive detection of intrusion behavior. The system dynamically adjusts the alarm trigger threshold based on changes in the coupled signal characteristics, thereby improving detection accuracy and false alarm resistance under different environmental interference conditions. Existing technologies of this type typically include the following steps: First, a complementary sensing channel is physically constructed through a dual-cable structure, enabling the two cables to generate mutually influential signal coupling characteristics when receiving external disturbances. Second, the system collects the raw signal data from the two cables, performs feature extraction and comparative analysis, and then constructs a coupling feature model. Next, an initial alarm threshold is set based on the model, and the threshold is corrected in real time through environmental sensing modules (such as temperature, humidity, and wind sensors) or historical data feedback mechanisms, forming a dynamic threshold update strategy. Finally, the dual-cable data is matched with the dynamic threshold to identify the existence of illegal intrusion behavior and trigger an alarm. The key to this type of method lies in extracting more discriminative disturbance features by utilizing the coupling relationship between the two cables, while effectively responding to the impact of environmental changes through a dynamic threshold adjustment mechanism, thereby improving the robustness and practicality of the system.

[0003] The existing technology has the following shortcomings:

[0004] In the process of dynamically determining the perimeter alarm threshold of a detection cable based on dual-cable coupling, when the detection cable passes through a metal pipe structure with electromagnetic reflection characteristics, the reflection, folding, and local shielding effects of the metal pipe on the electromagnetic signal cause the coupled signal generated by the same disturbance event to undergo multiple reflections and path differences during propagation. This results in response phenomena such as spectral misalignment, phase lag, or multi-peak superposition between the two cables. Existing dynamic determination technologies for perimeter alarm thresholds based on dual-cable coupling typically rely on the synchronization characteristics of the coupled signals in the frequency domain for matching and judgment. They lack a modeling and adaptation mechanism for the spectral misalignment characteristics of the signal in the aforementioned reflection structure. Therefore, they cannot accurately determine whether the disturbance reaches the triggering standard corresponding to the perimeter alarm threshold of the detection cable based on the spectral misalignment characteristics of the coupled signals when the detection cable passes through an electromagnetic reflection structure. This problem will cause the system to continuously misjudge such real disturbances as non-uniform disturbances, refusing to update the alarm threshold or trigger an alarm, thus forming a stable missed detection blind zone in such structural areas, seriously affecting the detection accuracy and reliability of the entire perimeter alarm system.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide a dynamic determination method for perimeter alarm threshold of detection cables based on dual-cable coupling, so as to solve the problems in the background art mentioned above.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a dynamic determination method for perimeter alarm threshold of detection cables based on dual-cable coupling, specifically including the following steps:

[0008] S1. After the disturbance occurs in the detection cable, the original response data of the dual-cable coupling signal is extracted, and a multi-time period spectrum sequence is generated through time-frequency joint analysis to construct a spectrum jump criterion set to reflect the starting point of the coupling anomaly.

[0009] S2. Based on the spectrum jump criterion set, calculate the consistency curvature of the scatter plot of the main frequency of the dual-cable coupled signal, analyze the periodic repeat term density and energy return ratio, and determine whether the detection cable passes through the electromagnetic reflection structure.

[0010] S3. In the case where the detection cable passes through the electromagnetic reflection structure, extract the relative slip distribution of the signal channels corresponding to the first cable and the second cable on the frequency axis in the dual-cable coupling signal, and generate a set of feature vectors to describe the spectral misalignment characteristics of the dual-cable coupling signal.

[0011] S4. Map the feature vector set to the preset disturbance confidence map area, calculate the projection ratio and dynamic centroid offset in each early warning response domain, and determine whether the disturbance reaches the triggering standard corresponding to the perimeter alarm threshold of the detection cable.

[0012] S5. Based on the judgment result, perform dynamic adjustment of the alarm threshold, and actively lower, gradually maintain, or keep the alarm threshold constant for different disturbance confidence levels.

[0013] Preferably, S1 is as follows:

[0014] After a disturbance occurs in the detection cable, the raw response data of the dual-cable coupling signal of the first and second cables are collected simultaneously. Based on a fixed time window, the response data is normalized in amplitude, reconstructed in time axis and aligned with the data start point to obtain a continuous signal segment for spectrum analysis.

[0015] The raw response data of the dual-cable coupled signal is input into the time-frequency joint analysis process, and short-time Fourier transform and continuous wavelet transform are executed in sequence to extract the main frequency path, amplitude distribution and phase change information in multiple time periods, and generate the corresponding multi-time period spectrum sequence.

[0016] In a multi-time-segment spectral sequence, the location where the rate of change of the main frequency trajectory on the frequency axis occurs is identified, and the frequency increment, amplitude difference and phase drift at the corresponding time point are extracted to construct a set of spectral jump criteria to reflect the starting point of coupling anomalies.

[0017] Preferably, S2 specifically includes the following steps:

[0018] S201. Based on the spectrum jump criterion set, extract the main frequency scatter point positions corresponding to each time index, connect them in time order to form the main frequency scatter point trajectory of the dual-cable coupled signal, and calculate the offset distance and offset direction change of adjacent scatter points on the frequency axis respectively. By performing cumulative smoothing processing on the continuous offset change, generate a consistent curvature to describe the overall bending degree of the main frequency scatter point trajectory.

[0019] S202. After completing the uniform curvature calculation, along the frequency interval corresponding to the main frequency scatter point trajectory, the number of times the same frequency component appears repeatedly in different time indices is counted to obtain the periodic repeat term density, and the energy in the return direction and the energy in the forward propagation direction in the spectrum are integrated and accumulated to calculate the energy return ratio.

[0020] S203. After obtaining the uniform curvature, periodic repetition density and energy return ratio, the three are compared with the pre-set structural judgment intervals one by one. When the uniform curvature is in the high bending interval and the periodic repetition density and energy return ratio both exceed the corresponding threshold, it is determined that the detection cable has crossed the electromagnetic reflection structure.

[0021] Preferably, S201 specifically refers to:

[0022] The positions of the main frequency scatter points corresponding to each time index are read from the spectrum jump criterion set, the main frequency scatter points are arranged in chronological order, and a continuous trajectory of the main frequency scatter points of the dual-cable coupled signal is constructed based on the connection between adjacent time indices.

[0023] The offset distance of adjacent main frequency scatter points on the frequency axis is calculated point by point along the main frequency scatter point trajectory, and the directional change between adjacent offset directions is calculated. The offset distance and directional change are combined to form a continuous offset change sequence.

[0024] The continuous offset variation sequence is segmented and accumulated, and a smoothing constraint is introduced during the accumulation process to suppress local discrete fluctuations. Based on the accumulation results, a consistent curvature is generated to characterize the overall bending degree of the main frequency scatter point trajectory.

[0025] Preferably, S3 is as follows:

[0026] In the case where the detection cable passes through the electromagnetic reflection structure, the response data of the signal channels corresponding to the first cable and the second cable on the frequency axis are extracted from the dual-cable coupling signal. The spectral amplitude trajectory and phase trajectory of the corresponding frequency band are extracted according to the time index alignment method, which are used to establish the frequency correspondence between the first cable and the second cable under the same time index.

[0027] By comparing the dominant frequency position and phase center position of the signal channels of the first cable and the second cable at each time index along the frequency axis, the frequency shift value and phase shift amount between each pair of response frequency points are calculated to form the relative shift distribution of the signal channels of the first cable and the second cable on the frequency axis.

[0028] Based on the relative slip distribution of the signal channels of the first and second cables on the frequency axis, the slip abrupt change point, slip direction inflection point and slip duration interval are extracted. A vector set containing the main frequency slip, slip interval length, slip point density and phase slip amplitude is constructed to generate a feature vector set to describe the spectral misalignment characteristics of the dual-cable coupled signal.

[0029] Preferably, S4 specifically includes the following steps:

[0030] S401. Standardize the feature vector set using a unified feature scale, and map the standardized feature vector set to a preset perturbation confidence map area. Divide multiple early warning response domains within the preset perturbation confidence map area according to the pre-defined response level boundaries, and assign the mapping coordinates corresponding to each feature vector to the corresponding early warning response domain.

[0031] S402. Statistically analyze the distribution of all feature vectors in each early warning response domain, calculate the ratio between the number of feature vectors in each early warning response domain and the total number of feature vectors, obtain the projection ratio in each early warning response domain, and calculate the weighted dynamic centroid position based on the coordinate distribution of all feature vectors in the disturbance confidence map area, and determine the degree of offset of the dynamic centroid relative to the reference center of the disturbance confidence map area.

[0032] S403. Compare the projection ratio and dynamic center of gravity offset within the highest warning level response domain with the triggering criteria corresponding to the perimeter alarm threshold of the detection cable. When the projection ratio within the highest warning level response domain exceeds the corresponding ratio threshold and the dynamic center of gravity offset exceeds the corresponding offset threshold, it is determined that the disturbance has reached the triggering criteria corresponding to the perimeter alarm threshold of the detection cable.

[0033] Preferably, S402 is as follows:

[0034] Compare the mapping coordinates of all feature vectors within the preset perturbation confidence map region with the boundary range of each early warning response domain, count the number of feature vectors contained in each early warning response domain, and calculate the ratio of this number to the total number of all feature vectors to obtain the projection ratio corresponding to each early warning response domain.

[0035] Based on the two-dimensional coordinate positions of all feature vectors in the preset perturbation confidence map region, a weighted coordinate set with feature intensity as the weighting factor is constructed, and the weighted average of all weighted coordinates is calculated to obtain the weighted dynamic centroid position coordinates of the current perturbation event.

[0036] Select the coordinates of the reference center point within the preset perturbation confidence map area, calculate the Euclidean distance between the weighted dynamic centroid position coordinates and the reference center point coordinates, and use this distance as the degree of dynamic centroid offset of the current perturbation event.

[0037] Preferably, S5 is as follows:

[0038] Based on the judgment results, the projection ratio and dynamic centroid shift of the feature vector set in the preset perturbation credibility map area are extracted. Based on the positional relationship between the two in the corresponding threshold, a perturbation credibility level classification model is constructed to classify the current perturbation event into high perturbation credibility level, medium perturbation credibility level or low perturbation credibility level.

[0039] After obtaining the disturbance confidence level, the alarm threshold is dynamically adjusted: when the disturbance confidence level is high, the alarm threshold is actively lowered; when the disturbance confidence level is medium, the alarm threshold is gradually maintained, and the alarm threshold is smoothly adjusted within the preset adjustment period using linear interpolation.

[0040] When the disturbance confidence level is low, the alarm threshold is kept constant, maintaining the alarm threshold value unchanged within the current period to prevent unnecessary alarms caused by low-confidence disturbances, thereby achieving dynamic adjustment of the alarm threshold under different disturbance confidence levels.

[0041] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0042] 1. This invention constructs a complete dual-cable coupled signal processing link, combining spectral jump criterion extraction, consistency curvature analysis, spectral misalignment modeling, and disturbance reliability level determination. This enables the identification of dominant frequency misalignment and phase anomalies caused by the detection cable traversing electromagnetic reflection structures. Compared to traditional techniques relying on frequency domain synchronization, this solution establishes for the first time a feature vector set and reliable spectrum mapping mechanism for coupled spectral misalignment. This allows the system to accurately determine whether disturbances meet alarm triggering criteria even when facing complex propagation characteristics such as multi-peaked spectra, phase drift, and path reversal, eliminating monitoring blind spots caused by structural obstruction. The introduction of periodic repeatability and energy return ratio further enhances the ability to identify interference modes caused by electromagnetic reflection and improves the adaptability to abnormal signal propagation structures.

[0043] 2. This invention introduces a disturbance credibility level classification model after disturbance determination. By dividing disturbance events into high, medium, and low credibility levels, it dynamically adjusts the alarm threshold execution strategy to achieve proactive lowering, gradual maintenance, or constant maintenance of the alarm threshold, constructing an alarm parameter adjustment mechanism with self-learning and adaptive capabilities. This mechanism enables the system to ensure high sensitivity in responding to real disturbances while avoiding false alarms caused by weak interference or mismatched signals, thereby improving detection accuracy while ensuring the stability and robustness of system operation. Overall, this technical solution improves the judgment accuracy and response reliability of dual-cable coupled detection systems in complex structural environments, providing an efficient and robust solution for perimeter safety monitoring of cables crossing reflective structures. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0045] Figure 1 is a flowchart illustrating the dynamic determination method for perimeter alarm threshold of detection cable based on dual-cable coupling according to the present invention. Detailed Implementation

[0046] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0047] This invention provides a method for dynamically determining the perimeter alarm threshold of a detection cable based on dual-cable coupling, as shown in Figure 1, which specifically includes the following steps:

[0048] S1. After the disturbance occurs in the detection cable, the original response data of the dual-cable coupling signal is extracted, and a multi-time period spectrum sequence is generated through time-frequency joint analysis to construct a spectrum jump criterion set to reflect the starting point of the coupling anomaly.

[0049] In this embodiment, S1 specifically refers to:

[0050] After a disturbance occurs in the detection cable, the raw response data of the dual-cable coupling signal of the first and second cables are collected simultaneously. Based on a fixed time window, the response data is normalized in amplitude, reconstructed in time axis and aligned with the data start point to obtain a continuous signal segment for spectrum analysis.

[0051] After a disturbance occurs in the detection cable, both the first and second cables will generate electromagnetic response signals to the same disturbance event. Collecting these two sets of signals separately is to establish the coupling relationship between the two cables. By synchronously sampling the two signal channels, the raw response data of the coupled signals can be obtained. During data extraction, each signal segment needs to be truncated using a fixed time window to ensure that each data segment has the same length and a unified starting point, facilitating subsequent comparison in the frequency domain. Amplitude normalization is used to eliminate level offsets between the two cables caused by environmental noise or differences in laying tension, ensuring that the signal amplitude is in a uniform dimension. Time axis reconstruction is based on the disturbance trigger point to rebuild the time index, ensuring that the signals of the two channels unfold under the same reference frame. Data start point alignment synchronizes the first response waveforms of the two channels, making the interference information comparable between the two channels. For example, when a disturbance generates a signal response, the first cable has a rising edge at time T1, and the second cable responds at T2. Through alignment, the two rising edge time points are synchronized to a unified reference T0, thereby avoiding spectral deviation. After completing the above processing, a continuous signal segment can be obtained for spectrum analysis, ensuring that the signal has good time-domain and amplitude comparability, which facilitates the extraction of spectrum features in the next stage.

[0052] Disturbance in the sensing cable refers to the sensing cable installed on the physical boundary of the perimeter being subjected to external mechanical disturbances, electromagnetic disturbances, or vibration interference, causing a change in its transmitted signal response. The first and second cables are two sensing lines laid parallel along the same boundary, coupled through electromagnetic fields or acoustic paths. The raw response data of the dual-cable coupled signal refers to the original electrical signal sequences generated by the first and second cables respectively after the disturbance occurs, used to characterize the response characteristics of the disturbance on both channels. A fixed time window refers to extracting continuous time intervals of data of a fixed length with the disturbance's starting point as a reference, ensuring a consistent time dimension between different sampling segments. Amplitude normalization is a linear transformation process that scales each signal segment to a uniform amplitude range, thereby eliminating comparison bias caused by differences in channel gain. Time axis reconstruction refers to re-establishing a unified time reference after identifying the disturbance's starting point to ensure a consistent temporal framework for the dual-channel signals during analysis. Data start-point alignment processing involves detecting the first response feature point (such as a rising edge or energy burst point) in the two signals using an algorithm and aligning it on the time axis to avoid phase drift caused by starting-point deviations in subsequent analysis. The continuous signal segment used for spectrum analysis refers to the set of time-domain signals that have been normalized and aligned to meet the requirements of subsequent frequency domain transformation. It is usually used as the input basis for subsequent short-time Fourier transform or wavelet analysis.

[0053] The raw response data of the dual-cable coupled signal is input into the time-frequency joint analysis process, and short-time Fourier transform and continuous wavelet transform are executed in sequence to extract the main frequency path, amplitude distribution and phase change information in multiple time periods, and generate the corresponding multi-time period spectrum sequence.

[0054] To accurately capture the dynamic characteristics of the disturbed signal in both the time and frequency domains, the raw response data of the dual-cable coupled signal needs to be input into a joint time-frequency analysis process. This process sequentially executes a short-time Fourier transform (SFT) and a continuous wavelet transform. The SFT is used to observe the spectral evolution of the signal in segments within different time windows, providing moderate time and frequency resolution to capture the changes in the dominant frequency distribution of sudden disturbances. The continuous wavelet transform supplements the local feature extraction capabilities of frequency and time at a finer granular level, making it suitable for processing non-stationary components in coupled signals. This joint process first divides the raw response data into windows, then performs a SFT on each window to obtain a two-dimensional spectrum of frequency changes over time. Then, a continuous wavelet transform is performed on the same signal to extract the detailed frequency response of local abrupt changes. Finally, by extracting the dominant frequency path changes, amplitude energy distribution, and phase drift characteristics on the frequency axis within each time period, a set of time-stamped spectral sequences is constructed. For example, a disturbance event may initially exhibit a concentrated dominant frequency of 5kHz, which later shifts to 6.2kHz due to changes in the cable structure. Using only a single analysis method would make it difficult to reconstruct this change process, while joint analysis can fully reveal the process.

[0055] The raw response data of a dual-cable coupled signal refers to continuous time-series data collected from the first and second cables, respectively. This data carries the initial electromagnetic response characteristics of the coupled signal after the disturbance occurs. The time-frequency joint analysis process is a combined analysis path that includes multiple spectrum extraction tools to simultaneously analyze the time and frequency characteristics of the signal, addressing the limitation of identifying non-stationary disturbances in single-dimensional analysis. The short-time Fourier transform is a method that divides the original signal into fixed-length sliding windows and performs Fourier operations within each window, primarily used to reveal the macroscopic trend of the signal's dominant frequency changing over time. The continuous wavelet transform is a multi-scale analysis method based on scalable and translational mother wavelets, capable of reconstructing details of short-term drastic changes or low-frequency gradual changes in the signal at different resolutions. The dominant frequency path within multiple time periods refers to the continuous trajectory of the maximum energy frequency points identified in different time windows, used to observe how the disturbance frequency evolves over time. Amplitude distribution refers to the response intensity of each frequency point in the time dimension, used to characterize the range of energy concentration. Phase change information is used to identify the phase shift trend in the coupled signal due to differences in propagation paths. Multi-time period spectrum sequences are data sets composed of the above types of information and arranged in an orderly manner, forming the core foundation for analyzing the evolution trend of disturbance characteristics.

[0056] In a multi-time-segment spectral sequence, the location where the rate of change of the main frequency trajectory on the frequency axis occurs is identified, and the frequency increment, amplitude difference and phase drift at the corresponding time point are extracted to construct a set of spectral jump criteria to reflect the starting point of coupling anomalies.

[0057] After obtaining the spectral sequence across multiple time periods, it is necessary to continuously track the dominant frequency position within each time period along the frequency axis to form a dominant frequency trajectory. By comparing the rate of change of the dominant frequency position in adjacent time periods, the location where the rate of change of the dominant frequency trajectory occurs can be identified. This location usually corresponds to the time point when the propagation conditions or response state of the coupled signal change. In practice, the frequency point with the highest energy proportion in each time period can be selected as the dominant frequency marker. Then, differential analysis is performed on the dominant frequency markers of consecutive time periods to obtain the frequency increment. At the same time, the amplitude value corresponding to the dominant frequency is extracted at the same time point and compared with the previous time period to obtain the amplitude difference. Further analysis is then performed on the phase change trend of the dual-cable signal at this frequency point to obtain the phase drift. For example, during a disturbance, the dominant frequency trajectory originally changed smoothly. When the signal propagation path changes due to the change in coupling conditions, the movement speed of the dominant frequency in adjacent time periods increases significantly. At the same time, the amplitude distribution shifts and the phase relationship deflects. By simultaneously satisfying the combined characteristics of frequency increment, amplitude difference, and phase drift, this time point can be determined as the starting point of coupling anomaly, and the relevant parameters are recorded to construct a spectral jump criterion set. This approach avoids misjudgment problems caused by relying on a single feature.

[0058] The location on the frequency axis where the rate of change of the dominant frequency trajectory occurs refers to the time point in a multi-time-segment spectral sequence where the trajectory of the dominant frequency evolution over time changes from a steady change to an accelerated or decelerated change. This point reflects a critical location where the signal propagation or coupling state changes. Frequency increment describes the displacement amplitude of the dominant frequency between adjacent time segments and is an important quantitative indicator for measuring the degree of frequency change. Amplitude difference represents the change in energy intensity of the same dominant frequency in different time segments, used to characterize whether the signal energy distribution has undergone a significant adjustment. Phase shift reflects the degree of shift in the phase relationship of the signal during propagation and is often used to identify phase response differences caused by changes in the propagation path or coupling relationship. The spectral jump criterion set used to reflect the starting point of coupling anomalies is a data set formed by combining frequency increment, amplitude difference, and phase shift in chronological order. This set centrally describes the spectral characteristic changes of the coupled signal when it transitions from a normal evolution state to an abnormal state, providing a foundation for subsequent determination of whether coupling anomalies are related to real disturbances.

[0059] S2. Based on the spectrum jump criterion set, calculate the consistency curvature of the scatter plot of the main frequency of the dual-cable coupled signal, analyze the periodic repeat term density and energy return ratio, and determine whether the detection cable passes through the electromagnetic reflection structure.

[0060] In this embodiment, S2 specifically includes the following steps:

[0061] S201. Based on the spectrum jump criterion set, extract the main frequency scatter point positions corresponding to each time index, connect them in time order to form the main frequency scatter point trajectory of the dual-cable coupled signal, and calculate the offset distance and offset direction change of adjacent scatter points on the frequency axis respectively. By performing cumulative smoothing processing on the continuous offset change, generate a consistent curvature to describe the overall bending degree of the main frequency scatter point trajectory.

[0062] S202. After completing the uniform curvature calculation, along the frequency interval corresponding to the main frequency scatter point trajectory, the number of times the same frequency component appears repeatedly in different time indices is counted to obtain the periodic repeat term density, and the energy in the return direction and the energy in the forward propagation direction in the spectrum are integrated and accumulated to calculate the energy return ratio.

[0063] To further identify the characteristics of the probe cable traversing the electromagnetic reflection structure after completing the uniform curvature calculation, it is necessary to conduct periodicity and energy characteristic analysis along the frequency range corresponding to the dominant frequency scatter plot. Specifically, firstly, all frequency points are extracted from the constructed dominant frequency scatter plot, and the repetition frequency of these frequency points is counted at different time indices using a sliding window approach. By setting the time overlap threshold and frequency matching accuracy, frequency components that repeatedly appear at multiple times can be identified. These repetitions usually originate from multiple reflections of electromagnetic waves within the metal pipe, thus forming a periodic repetitive response. After standardizing the repetition count of each frequency component, a periodic repetition density can be constructed to measure the intensity of periodic reflection phenomena in the signal. Next, to analyze the energy characteristics of reflection intensity and propagation direction, it is necessary to integrate and accumulate the energy components propagating in the reverse direction and the energy components propagating in the forward direction in the spectrum. This operation can separate the energy flow based on the phase steering diagram, and then sum the components according to the frequency and time indices to obtain the overall return energy and forward energy values. The final ratio between the two is the energy return ratio.

[0064] The frequency range corresponding to the dominant frequency scatter plot refers to the range covered by the scatter plot extending along the time axis in the frequency dimension. This frequency range centrally embodies the most characteristic dominant frequency response in the dual-cable coupled signal. The same frequency component refers to frequency points that repeat in different time indices, usually indicating periodic interference or structural reflection echoes. The periodic repetition density is a normalized measure composed of the frequency of occurrence of each frequency point in the time dimension, reflecting whether the frequency is the dominant frequency signal of stable return interference. The division of energy in the return direction and energy in the forward propagation direction in the spectrum depends on the judgment of the signal phase change trend. Forward propagation signals generally have a linearly increasing phase, while backward propagation signals exhibit a reverse drift. Integral accumulation is a numerical accumulation operation of energy values ​​in the continuous time and frequency dimensions, used to quantify the total energy distribution characteristics. The energy return ratio, as the ratio of the two types of energy, reflects the significance of the reflection component in the signal and is one of the important criteria for determining the existence of electromagnetic reflection structures.

[0065] S203. After obtaining the uniform curvature, periodic repetition density and energy return ratio, the three are compared with the pre-set structural judgment intervals one by one. When the uniform curvature is in the high bending interval and the periodic repetition density and energy return ratio both exceed the corresponding threshold, it is determined that the detection cable has crossed the electromagnetic reflection structure.

[0066] To determine whether a probe cable has traversed an electromagnetic reflection structure, the uniform curvature, periodic repetition density, and energy return ratio need to be precisely compared sequentially. First, historical signal data under typical electromagnetic reflection environments are pre-collected and calibrated to construct a structure determination interval encompassing high bending characteristics, uniform periodic repetition response, and a high energy return ratio. This interval sets one or more reference thresholds for each criterion using a statistical distribution. In practical applications, the uniform curvature, periodic repetition density, and energy return ratio calculated for the current disturbance are compared item by item with the corresponding thresholds in the structure determination interval. Only when the uniform curvature reaches the preset high bending interval, and the periodic repetition density and energy return ratio both exceed their respective lower thresholds, is it determined that a probe cable has traversed an electromagnetic reflection structure. This joint determination method significantly improves the identification accuracy and effectively eliminates interference from non-structural disturbances.

[0067] The pre-defined structural judgment interval is a feature judgment range extracted and statistically analyzed from a large amount of experimental data, used to reflect the typical variation patterns of signal parameters in an electromagnetic reflection environment. Item-by-item comparison refers to comparing the three parameter values ​​to be judged one by one with the reference thresholds within the structural judgment interval to determine whether the judgment conditions are met simultaneously. The high bending interval is a numerical range set for consistent curvature, typically corresponding to situations where the dominant frequency scatter point trajectory exhibits severe spatial reversals. The corresponding threshold for the periodic repetition density is used to screen for strong periodic repetition components in the spectrum, while the corresponding threshold for the energy backpropagation ratio reflects the proportion of backpropagation components in the energy distribution. The joint exceedance of these three indicators can effectively identify atypical coupling signal behavior generated by the probe cable in the reflection structure, serving as a key criterion for further modeling or classification.

[0068] In this embodiment, S201 specifically refers to:

[0069] The positions of the main frequency scatter points corresponding to each time index are read from the spectrum jump criterion set, the main frequency scatter points are arranged in chronological order, and a continuous trajectory of the main frequency scatter points of the dual-cable coupled signal is constructed based on the connection between adjacent time indices.

[0070] When reading the dominant frequency scatter points corresponding to each time index in the spectral jump criterion set, each time index represents a specific signal observation interval within a time window, and the dominant frequency scatter point position refers to the frequency point with the highest energy in the spectrum of the dual-cable coupled signal within that time window. This position is obtained by scanning the frequency-energy distribution in the short-time Fourier transform or wavelet transform results. To construct a continuous trajectory reflecting the frequency change trend, the dominant frequency scatter points corresponding to all time indices need to be arranged in chronological order, i.e., generating a frequency point sequence that increases in time. Subsequently, using the dominant frequency scatter points between adjacent time indices as connection points, they are connected in chronological order along the frequency dimension to draw a continuous dominant frequency scatter point trajectory of the dual-cable coupled signal. This trajectory reflects the trend of the dominant frequency evolving over time under disturbance and is the core basis for identifying abnormal frequency fluctuations in the reflection region. For example, when the probe cable crosses an electromagnetic reflection structure, this trajectory will exhibit characteristics such as nonlinear backtracking and drastic frequency jumps. By constructing this continuous trajectory, the foundation can be laid for subsequent calculation of uniform curvature and identification of electromagnetic reflection regions. During execution, to ensure the continuity and accuracy of the trajectory, a frequency jump threshold can be introduced to filter discrete outliers, thereby improving the smoothness and stability of the main frequency trajectory.

[0071] The offset distance of adjacent main frequency scatter points on the frequency axis is calculated point by point along the main frequency scatter point trajectory, and the directional change between adjacent offset directions is calculated. The offset distance and directional change are combined to form a continuous offset change sequence.

[0072] In the constructed scatter plot of the dominant frequency of the dual-cable coupled signal, each dominant frequency point corresponds to a time index and its dominant frequency value. By traversing adjacent points in the trajectory, their offset distance on the frequency axis can be calculated pairwise, i.e., the absolute value of the frequency value of the subsequent point minus the absolute value of the frequency value of the preceding point. This offset distance reflects the drastic change in the dominant frequency. Simultaneously, the offset direction must be recorded, i.e., whether the dominant frequency value is rising, falling, or remaining constant. Between consecutive points, changes in the offset direction are compared; for example, a change from rising to falling or from falling to rising is marked as a sudden change in direction. Combining the offset distance of each pair of adjacent points with the corresponding directional change constructs a complete continuous offset change sequence, which can be used to capture abnormal fluctuation patterns in the dominant frequency trajectory. This method can be used to identify signal reflection phenomena caused by electromagnetic reflection structures. For example, in the reflection-affected region, the offset distance between adjacent dominant frequency points is usually large, and the direction changes frequently, manifesting as large jumps in high frequencies in the continuous offset change sequence. This sequence provides the necessary raw data support for subsequent calculation of the uniform curvature, which helps to quantify the overall smoothness or bending characteristics of the signal's main frequency change trajectory, thereby assisting in the identification of abnormal structural segments.

[0073] The continuous offset variation sequence is segmented and accumulated, and a smoothing constraint is introduced during the accumulation process to suppress local discrete fluctuations. Based on the accumulation results, a consistent curvature is generated to characterize the overall bending degree of the main frequency scatter point trajectory.

[0074] Segmented accumulation processing of continuous offset sequences involves dividing the entire sequence into several fixed-length continuous time intervals according to time sequence, and accumulating the changes in offset distance and direction within each interval to extract the overall offset trend of that segment. This processing method effectively preserves the continuity characteristics of each segment and reduces the interference of minor disturbances on the overall judgment. To further suppress drastic fluctuations caused by environmental noise or instantaneous disturbances in certain time periods, smoothing constraints need to be introduced during the accumulation process. Smoothing constraints are usually implemented through moving averages or Gaussian weighting, ensuring that each segmented accumulation result has good stability while maintaining trend changes, reducing misjudgments caused by individual outliers. Local discrete fluctuations refer to rapid, discontinuous changes in the trajectory. These fluctuations often originate from unstructured disturbances and lack stable reflection characteristics, thus requiring smoothing processing to improve the accuracy of subsequent judgments. Finally, by analyzing the curvature change trend between the accumulated values ​​of each segment—that is, using the growth direction and amplitude of adjacent accumulated segments as input—the continuous curvature is calculated to generate a consistent curvature parameter. This parameter characterizes the degree of curvature of the dominant frequency scatter trajectory in the overall space. A large uniform curvature indicates that the trajectory exhibits strong nonlinear changes, often corresponding to situations where the detection cable crosses complex structural areas such as metal reflective pipe sections. This parameter provides crucial support for subsequent determination of the presence of electromagnetic reflection structures.

[0075] S3. In the case where the detection cable passes through the electromagnetic reflection structure, extract the relative slip distribution of the signal channels corresponding to the first cable and the second cable on the frequency axis in the dual-cable coupling signal, and generate a set of feature vectors to describe the spectral misalignment characteristics of the dual-cable coupling signal.

[0076] In this embodiment, S3 specifically refers to:

[0077] In the case where the detection cable passes through the electromagnetic reflection structure, the response data of the signal channels corresponding to the first cable and the second cable on the frequency axis are extracted from the dual-cable coupling signal. The spectral amplitude trajectory and phase trajectory of the corresponding frequency band are extracted according to the time index alignment method, which are used to establish the frequency correspondence between the first cable and the second cable under the same time index.

[0078] In cases where a detection cable traverses an electromagnetic reflection structure, it is necessary to extract the frequency response data of the dual-cable coupled signal channels corresponding to the first and second cables. This can be achieved by performing short-time Fourier transforms on both signals to obtain their spectral images at each time index, and then extracting the spectral amplitude and phase trajectories within the corresponding frequency bands. To ensure the accuracy of the analysis, a time index alignment method is required to maintain strict time synchronization for each frame of data from the dual-cable signals, ensuring that the frequency response comparison is based on the responses triggered by the same physical event. Based on this, by comparing the dominant frequency position and phase centroid of the corresponding frequency bands, the frequency correspondence between the first and second cable signals at the same time index can be established, providing a precise benchmark for subsequent slip analysis.

[0079] "The frequency response data of the signal channels corresponding to the first and second cables in the dual-cable coupling signal" refers to the frequency domain response information collected by two parallel sensing cables for the same disturbance event. "The spectral amplitude and phase trajectories of the corresponding frequency bands" are tracking curves of energy distribution and phase changes at different time points within the same frequency range, reflecting the propagation characteristics of electromagnetic disturbances. "The frequency correspondence between the first and second cables under the same time index" refers to comparing the frequency responses of the two cables frequency by frequency at the same moment to identify coupling differences such as whether the main peak of the spectrum is misaligned or the phase center is shifted, for further analysis of the spatial distribution and dynamic characteristics of the spectral misalignment. By constructing these correspondences, frequency slippage caused by electromagnetic reflection can be identified more accurately, providing core criteria for determining electromagnetic anomalies.

[0080] By comparing the dominant frequency position and phase center position of the signal channels of the first cable and the second cable at each time index along the frequency axis, the frequency shift value and phase shift amount between each pair of response frequency points are calculated to form the relative shift distribution of the signal channels of the first cable and the second cable on the frequency axis.

[0081] To extract the relative slip distribution of the first and second cable signal channels on the frequency axis, the frequency domain response of the two channels needs to be analyzed synchronously at each time index. Specifically, the dominant frequency position and phase center position of the two channels are extracted at each time point. The dominant frequency position can be obtained by locating the frequency point of maximum amplitude, and the phase center position can be calculated based on a weighted phase average. Then, comparisons are performed point-by-point along the entire frequency axis. For the difference in the dominant frequency positions of the two channels within the same frequency band, the frequency slip value is calculated; then, the phase slip amount is calculated by comparing the differences between the phase center positions. The frequency slip values ​​and phase slip amounts at all time indices constitute a continuous two-dimensional distribution sequence, i.e., the relative slip distribution of the first and second cable signal channels on the frequency axis, used to characterize the differences in coupling response between the cables. This processing method can accurately reflect the differences in propagation paths caused by electromagnetic reflection.

[0082] The dominant frequency position refers to the frequency point in the spectrum where energy is most concentrated, representing the main propagation component of the signal; the phase center position is the centroid of phase change within the frequency range, reflecting the average phase characteristics of the signal. Frequency slip represents the relative offset of the dominant frequency between two channels on the frequency axis, measured in Hertz; phase slip represents the phase center offset angle between two signals, measured in radians or degrees. These parameters form a time-varying slip map through time-sequential indexing analysis, called the relative slip distribution. This distribution not only reveals the dynamic evolution of spectral misalignment but can also be used to identify the impact of abnormal propagation paths or reflection interference on coupling characteristics, serving as the core basis for constructing spectral misalignment feature vectors.

[0083] Based on the relative slip distribution of the signal channels of the first and second cables on the frequency axis, the slip abrupt change point, slip direction inflection point and slip duration interval are extracted. A vector set containing the main frequency slip, slip interval length, slip point density and phase slip amplitude is constructed to generate a feature vector set to describe the spectral misalignment characteristics of the dual-cable coupled signal.

[0084] To generate a feature vector set describing the spectral misalignment characteristics of dual-cable coupled signals, key location points are first extracted based on the relative slip distribution of the signal channels of the first and second cables along the frequency axis. In the slip curve, points where the slip value increases or decreases drastically in a continuous time index are marked as slip abrupt change points; points where the slip value trend changes from positive to negative or inverse are identified as slip direction inflection points; and time intervals where the slip value remains within a specific direction or amplitude range are identified as slip duration intervals. Subsequently, within each slip duration interval, the dominant frequency slip is calculated as the slip intensity index for that segment, the time index span is statistically analyzed as the slip interval length, the frequency of abrupt change points or inflection points in that interval is quantified as the slip point density, and the total phase slip amplitude for each segment is evaluated. These values ​​are then used to construct an ordered vector set to comprehensively represent the spectral misalignment behavior of the dual-cable coupled signal. This feature vector set provides high-resolution description capabilities when dealing with coupling offsets caused by electromagnetic reflections, facilitating subsequent spectral projection and threshold determination.

[0085] A slip abrupt change point refers to the location where the frequency or phase slip value changes significantly within a short period of time, usually caused by abrupt changes in the propagation path or coupling reflection. A slip direction inflection point indicates a critical node where the slip trend changes directionally, reflecting propagation path reversal or interference reconstruction phenomena. The slip duration interval is the time period during which the slip trend remains consistent, characterizing the sustained impact range of a certain type of interference. The dominant frequency slip is the average slip amplitude between the dominant frequencies of the two cables; the slip interval length is the length of the slip duration; the slip point density refers to the frequency of slip anomalies per unit time; and the phase slip amplitude is the total phase drift. These technical characteristics are encoded through vector groups to form a multi-dimensional expression of coupling misalignment features, constituting a feature vector set that can be used to identify abnormal coupling situations and support dynamic determination of perimeter alarm thresholds for detection cables.

[0086] S4. Map the feature vector set to the preset disturbance confidence map area, calculate the projection ratio and dynamic centroid offset in each early warning response domain, and determine whether the disturbance reaches the triggering standard corresponding to the perimeter alarm threshold of the detection cable.

[0087] In this embodiment, S4 specifically includes the following steps:

[0088] S401. Standardize the feature vector set using a unified feature scale, and map the standardized feature vector set to a preset perturbation confidence map area. Divide multiple early warning response domains within the preset perturbation confidence map area according to the pre-defined response level boundaries, and assign the mapping coordinates corresponding to each feature vector to the corresponding early warning response domain.

[0089] Unified feature scaling is typically achieved by applying normalization or standard deviation standardization to each feature dimension, mapping features with different dimensions and numerical ranges to the same data scale space. For example, each feature value is converted to a normally distributed value with a mean of zero and a standard deviation of one. After standardization, the set of feature vectors can be mapped to a predefined perturbation confidence map region. This mapping process can use dimensionality reduction methods such as principal component analysis, linear discriminant analysis, or t-SNE to compress high-dimensional feature vectors into coordinate representations in two- or three-dimensional space. In the mapping result, the perturbation confidence map region is pre-divided into response level boundaries, such as "safe domain," "early warning domain," and "high-risk domain." The coordinates of each standardized feature vector obtained after dimensionality reduction mapping fall within the corresponding boundary, and the feature vector is assigned to the corresponding early warning response domain. The purpose of this process is to establish a mapping relationship between feature behavior and alarm level in the visualization map, facilitating subsequent measurement of the perturbation level.

[0090] Unified feature scaling refers to normalizing the dimensions of the original feature set using consistent statistics or rules to ensure that subsequent calculations are not distorted due to inconsistent dimensions. Standardization typically involves stretching and compressing feature values ​​based on the mean and standard deviation, giving each feature dimension the same statistical basis. A pre-constructed perturbation confidence map region is a pre-built map space to accommodate multiple types of perturbation features, visually reflecting the distribution of feature vectors under different perturbation scenarios. Pre-defined response level boundaries divide the map region into multiple areas according to perturbation risk levels, with each area corresponding to an alarm level. The early warning response domain is the set of areas defined by the level boundaries, used to classify the risk level of each perturbation sample from the feature vector set within the map space. This technical structure ensures that the spatial correspondence between perturbation features and alarm thresholds is clear, hierarchical, and measurable.

[0091] S402. Statistically analyze the distribution of all feature vectors in each early warning response domain, calculate the ratio between the number of feature vectors in each early warning response domain and the total number of feature vectors, obtain the projection ratio in each early warning response domain, and calculate the weighted dynamic centroid position based on the coordinate distribution of all feature vectors in the disturbance confidence map area, and determine the degree of offset of the dynamic centroid relative to the reference center of the disturbance confidence map area.

[0092] S403. Compare the projection ratio and dynamic center of gravity offset within the highest warning level response domain with the triggering criteria corresponding to the perimeter alarm threshold of the detection cable. When the projection ratio within the highest warning level response domain exceeds the corresponding ratio threshold and the dynamic center of gravity offset exceeds the corresponding offset threshold, it is determined that the disturbance has reached the triggering criteria corresponding to the perimeter alarm threshold of the detection cable.

[0093] To determine whether a disturbance has reached the triggering criteria corresponding to the perimeter alarm threshold for cable detection, the first step is to compare the projection proportion of the previously calculated feature vector within the highest warning level response domain with the degree of dynamic centroid shift. The comparison is based on pre-set dual threshold parameters of the cable detection perimeter alarm system, including a proportion threshold for judging the concentration of disturbance intensity distribution and a shift threshold for measuring the degree of disturbance focusing. When the projection proportion exceeds the proportion threshold, it indicates that the disturbance event is concentrated in a high-risk area in the feature map; simultaneously, if the degree of dynamic centroid shift exceeds the shift threshold, it indicates that the overall disturbance is significantly shifting towards a higher risk level. The system only determines that the current disturbance has reached the triggering criteria corresponding to the cable detection perimeter alarm threshold and triggers the corresponding response mechanism when both indicators simultaneously meet their respective threshold conditions. This process possesses the ability to jointly assess the disturbance level trend and distribution intensity, significantly reducing the risk of false alarms caused by local anomalies or feature outliers.

[0094] The highest warning level response domain refers to the characteristic distribution area representing the strongest intrusion risk level in the disturbance credibility map region, usually located at the map boundary or within the high-risk cluster zone. The triggering criteria corresponding to the perimeter alarm threshold of the detection cable are composed of multi-dimensional characteristic parameters and are the core reference for assessing whether the intensity of the disturbance event is sufficient to trigger an alarm. The corresponding proportion threshold sets the minimum judgment standard for the degree of feature clustering in the highest level response domain. The corresponding offset threshold defines the minimum offset amplitude required for the disturbance center of gravity to deviate from the baseline state. Together, these two constitute the threshold conditions of the multi-feature joint judgment mechanism, which helps to accurately identify real disturbance events and improve the judgment accuracy and robustness of the perimeter alarm system.

[0095] In this embodiment, S402 specifically refers to:

[0096] Compare the mapping coordinates of all feature vectors within the preset perturbation confidence map region with the boundary range of each early warning response domain, count the number of feature vectors contained in each early warning response domain, and calculate the ratio of this number to the total number of all feature vectors to obtain the projection ratio corresponding to each early warning response domain.

[0097] To obtain the projection percentage corresponding to each early warning response domain, it is necessary to first accurately match the mapping coordinates of all feature vectors in the perturbation confidence map region with the boundary range of the early warning response domain. The boundary range of the early warning response domain is usually predefined using geometric partitioning methods, such as using polygonal regions, elliptical partitioning, or regular grid division methods in a two-dimensional plane to divide the map into multiple sub-regions with hierarchical meaning. For the mapping coordinates of each feature vector, the domain assignment is determined by whether it falls inside a specific set of boundary coordinates. This can be achieved using the ray method where a point lies inside a polygon or by comparing the distance between a point and the center of the region with the radius. Each feature vector falling into different early warning response domains is classified and accumulated, and the number of feature vectors contained in each early warning response domain is counted. This count is then divided by the total number of all feature vectors to obtain the projection percentage corresponding to that early warning response domain. Taking three types of early warning response domains as an example, if 60% of the feature vectors corresponding to a certain perturbation event are distributed in the high-risk response domain, 30% in the medium-risk early warning response domain, and 10% in the safe response domain, then the intensity and spatial feature concentration trend of the current perturbation event can be assessed. Boundary extents are used to define the geometric boundaries of each response level within the perturbation credibility map region. Their accuracy and clarity directly determine the accuracy of subsequent classification. The construction of boundary extents typically references historical data cluster centers, high-dimensional space mapping projection density, or expert experience to ensure coverage of typical distribution areas in the perturbation sample space. Projection ratio refers to the proportion of feature vectors contained within each response domain relative to the total number of feature vectors. It measures the central tendency and distribution weight of perturbation samples at each response level and is a crucial metric for subsequent alarm judgment and threshold adaptation. This process forms the static distribution basis for perturbation credibility assessment.

[0098] Based on the two-dimensional coordinate positions of all feature vectors in the preset perturbation confidence map region, a weighted coordinate set with feature intensity as the weighting factor is constructed, and the weighted average of all weighted coordinates is calculated to obtain the weighted dynamic centroid position coordinates of the current perturbation event.

[0099] To obtain the weighted dynamic centroid coordinates of the current disturbance event within the disturbance confidence map region, it is first necessary to define the mapping coordinates of each eigenvector in the two-dimensional map and assign corresponding weights based on the eigenvector's corresponding eigenintensity, thus constructing a weighted coordinate set. The eigenintensity can be uniformly quantified based on key values ​​contained in the eigenvector, such as dominant frequency shift, phase shift amplitude, or shift point density, reflecting the importance and signal influence of each eigenvector in the overall disturbance response. When constructing the weighted coordinate set, each two-dimensional coordinate is multiplied by its corresponding eigenintensity as a weighting factor, and all weighted coordinate values ​​are summed. Finally, the result is divided by the sum of all eigenintensities to obtain the weighted dynamic centroid coordinates. These coordinates are used to characterize the concentration trend of the disturbance event's response features within the disturbance confidence map region, more accurately reflecting the dominant direction and degree of focus of the disturbance intensity. Feature intensity is a scalar value extracted from the eigenvector to quantify the influence of a certain disturbance component, serving as the basis for distinguishing between strong and weak disturbance signals. The weighting factor is the specific numerical representation of the feature intensity in coordinate weighting, used to enhance the influence of features with higher disturbance contributions in centroid calculation. The weighted coordinate set is the set obtained by fusing all two-dimensional mapped coordinates with their corresponding weighting factors point by point, reflecting the centroid composition of the disturbance event in the response spectrum. The weighted dynamic centroid position coordinates are the central tendency points of this set based on the weight distribution; the degree of spatial offset can be used to determine changes in the disturbance confidence level. The overall process ensures that centroid assessment is more sensitive and discriminative under different disturbance conditions.

[0100] Select the coordinates of the reference center point within the preset perturbation confidence map area, calculate the Euclidean distance between the weighted dynamic centroid position coordinates and the reference center point coordinates, and use this distance as the degree of dynamic centroid offset of the current perturbation event.

[0101] When determining the degree of dynamic centroid shift in a disturbance event, a reference center point coordinate system must first be selected within a pre-defined disturbance confidence map region. This reference center point coordinate system is typically set at the centroid of a region within the map region representing a low-confidence-level disturbance or a normal-state disturbance, serving as a stability benchmark. The weighted dynamic centroid coordinates calculated earlier are then compared with these reference center point coordinates. The difference components between the horizontal and vertical axes in the two-dimensional coordinate system are extracted, and the Euclidean distance between the two points is calculated based on geometric principles. This Euclidean distance represents the spatial offset of the weighted dynamic centroid coordinates relative to the reference center point coordinates. It measures the degree of deviation of the overall distribution of the current disturbance event under the dominance of feature intensity, and is thus used to assess the anomalousness of the disturbance level. A large offset often indicates a significant anomalous clustering trend in the disturbance features, potentially corresponding to a higher level of intrusion risk. The reference center point coordinates set within the preset perturbation confidence map area are a set of two-dimensional static coordinate values, representing the concentration location of the mean perturbation features under ideal conditions. Euclidean distance is a standard metric tool used to calculate the true straight-line distance between any two points in two-dimensional space, possessing good geometric interpretation and computational stability. The dynamic centroid offset is the representation of this distance in the perturbation map evaluation system, used to quantify the dispersion and focusing trend changes of perturbation events in the confidence map space, and is one of the core indicators for triggering alarm thresholds. The overall design can stably evaluate the perturbation intensity level under multi-feature fusion and support dynamic response strategy adjustments.

[0102] S5. Based on the judgment result, perform dynamic adjustment of the alarm threshold, and actively lower, gradually maintain, or keep the alarm threshold constant for different disturbance confidence levels.

[0103] In this embodiment, S5 specifically refers to:

[0104] Based on the judgment results, the projection ratio and dynamic centroid shift of the feature vector set in the preset perturbation credibility map area are extracted. Based on the positional relationship between the two in the corresponding threshold, a perturbation credibility level classification model is constructed to classify the current perturbation event into high perturbation credibility level, medium perturbation credibility level or low perturbation credibility level.

[0105] Effective classification of disturbance confidence levels can be achieved by extracting two key indicators from the feature vector set after mapping to a preset disturbance confidence map region: projection ratio and dynamic centroid offset. Specifically, firstly, the proportion of feature vectors falling into different warning response domains is statistically analyzed to obtain the projection ratio of each response domain. Secondly, the dynamic centroid position of all feature vectors in the map coordinate system is calculated based on feature intensity weighting, and the Euclidean distance between them and the reference center point is calculated to obtain the dynamic centroid offset. These two indicators are then compared with multiple pre-defined judgment thresholds, and mapped to the corresponding confidence level classification region based on their placement within the interval. For example, if the projection ratio exceeds the highest warning response domain threshold and the dynamic centroid offset is at the upper limit of the offset interval, it is classified as a high disturbance confidence level; if both indicators are in the middle range, it is classified as a medium disturbance confidence level; if neither reaches any trigger threshold, it is classified as a low disturbance confidence level. This process allows for dynamic response to differences in disturbance intensity and pattern, thus providing an accurate basis for subsequent alarm threshold adjustment.

[0106] The perturbation credibility level classification model is a multi-threshold judgment framework built based on the feature vector distribution characteristics in the perturbation spectrum. It is used to classify the current perturbation event into three mutually exclusive levels. A high perturbation credibility level indicates that the feature vectors are concentrated in the high-risk response region, and the dynamic centroid deviates significantly from the center point, typically indicating a high degree of consistency between the perturbation signal characteristics and historical alarm events. A medium perturbation credibility level corresponds to a relatively dispersed feature vector distribution, but the dynamic centroid deviates to some extent, reflecting that the perturbation may be an atypical risk event. A low perturbation credibility level characterizes the perturbation signal characteristics in the spectrum close to the background or reference region, without showing a significant concentration or shift trend, usually related to environmental perturbations or interference events. This three-part model enables more discriminative hierarchical responses to perturbation events, improving the overall system's adaptability and false alarm suppression effect.

[0107] After obtaining the disturbance confidence level, the alarm threshold is dynamically adjusted: when the disturbance confidence level is high, the alarm threshold is actively lowered to enhance the ability to identify subsequent disturbances in the current detection cable monitoring area; when the disturbance confidence level is medium, the alarm threshold is gradually maintained by using linear interpolation to smoothly adjust the alarm threshold within the preset adjustment period to maintain monitoring stability and adaptability.

[0108] Once the disturbance confidence level is obtained, to improve the monitoring system's adaptability and response accuracy to disturbances of different levels, dynamic adjustment of the alarm threshold can be implemented according to the disturbance confidence level. For high-confidence-level disturbances, the alarm threshold can be actively lowered directly by correcting the current alarm threshold from the original set value to a lower value, thereby enhancing the sensitivity to subsequent similar disturbances. For example, if the original threshold is 100 units, it can be directly reduced to 80 units under high-confidence-level conditions, enabling the system to respond more quickly to continuous disturbance trends. For medium-confidence-level disturbances, a gradual maintenance operation is required, that is, making appropriate adjustments while maintaining the current alarm capability. In this case, a linear interpolation method can be used to continuously adjust the threshold in small steps within a preset adjustment period, gradually transitioning it between the original value and the target equilibrium value. For example, it can slowly change from 100 units to 90 units within a 10-minute period to balance threshold stability and adaptability to subsequent disturbances. This dynamic response mechanism can effectively avoid false alarms caused by over-adjustment and can also quickly improve alertness in high-risk situations.

[0109] Active threshold reduction refers to immediately lowering the alarm threshold by a significant margin upon identifying a high level of disturbance confidence, making the alarm system more sensitive to subsequent disturbance signals. This is commonly used for rapid response to high-intensity intrusions or persistent signal anomalies. Gradual maintenance is a progressive control mechanism that avoids system fluctuations caused by abrupt adjustments, making it particularly suitable for handling moderate-intensity but potentially risky disturbances. Linear interpolation is a commonly used numerical smoothing method that generates a continuously changing sequence of intermediate values ​​between the initial alarm threshold and the target adjustment threshold, ensuring a smooth, continuous, and abrupt threshold change. A preset control period is a fixed time window set by the system based on environmental characteristics or historical response experience. It controls the time scale of alarm threshold adjustments, avoiding system instability caused by frequent adjustments and ensuring that the system completes threshold adaptation within a specific timeframe. The core objective of smooth alarm threshold adjustment is to enhance monitoring robustness, maintaining responsiveness to disturbance changes while reducing the probability of false alarms, thereby constructing a sensitive and stable dynamic monitoring system.

[0110] When the disturbance confidence level is low, the alarm threshold is kept constant, maintaining the alarm threshold value unchanged within the current period to prevent unnecessary alarms caused by low-confidence disturbances, thereby achieving dynamic adjustment of the alarm threshold under different disturbance confidence levels.

[0111] When a disturbance is classified as low, a constant alarm threshold maintenance operation is required to prevent the system from overreacting to weak, risk-free disturbances. This operation maintains the current alarm threshold value at a stable level, ensuring that the monitoring system will not falsely trigger alarms due to slight fluctuations within the current cycle. This can be achieved by temporarily freezing the threshold update channel, blocking any threshold adjustment requests based on low-level disturbances, and maintaining the predetermined threshold unchanged within a set time window. For example, if the alarm threshold is initially set to 100 units, after identifying a low-level disturbance, the control system will skip the threshold update logic and maintain the 100-unit value until a new disturbance analysis cycle begins. This mechanism is particularly suitable for scenarios with frequent background noise disturbances or environmental microwave excitation signals, significantly reducing false alarm frequency and improving the system's ability to identify critical anomalies.

[0112] Constant holding is a stability constraint in the dynamic adjustment strategy of alarm thresholds. Its core lies in statically locking the threshold parameter to prevent unnecessary fluctuations in system sensitivity caused by an underestimated disturbance level. This operation is typically executed automatically based on the disturbance confidence level judgment result, requiring no manual intervention. When the disturbance confidence level is low, it means that the feature vector of the current disturbance event has a small projection proportion and a low degree of dynamic centroid shift in the disturbance confidence map region, and the system determines that it is insufficient to constitute a security threat. Forcibly adjusting the threshold at this time may lead to an increase in the false alarm rate or waste of resources. Therefore, by constantly holding the alarm threshold, the interference of weak disturbance events on the alarm logic can be effectively isolated, maintaining the stable operation of the monitoring system and reserving sensitivity redundancy space for high-level disturbances. This operation is also one of the important guarantee mechanisms for ensuring the long-term online reliability of the system.

[0113] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions according to the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means (e.g., infrared, wireless, microwave, etc.). A computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. Available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. Semiconductor media can be solid-state drives.

[0114] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0115] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0116] In the several embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0117] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0118] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0119] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for dynamically determining the perimeter alarm threshold of a detection cable based on dual-cable coupling, characterized in that, Specifically, the following steps are included: S1. After the detection cable experiences a disturbance, extract the original response data of the dual-cable coupling signal, generate a multi-time-segment spectrum sequence through time-frequency joint analysis, and construct a spectrum jump criterion set to reflect the starting point of the coupling anomaly; S2. Based on the spectrum jump criterion set, calculate the consistency curvature of the scatter plot of the main frequency of the dual-cable coupling signal, analyze the periodic repeat term density and the energy return ratio, and determine whether the detection cable has crossed the electromagnetic reflection structure; S3. If the detection cable has crossed the electromagnetic reflection structure, extract the relative slip distribution of the signal channels corresponding to the first and second cables on the frequency axis in the dual-cable coupling signal, and generate a feature vector set to describe the spectral misalignment characteristics of the dual-cable coupling signal; S4. Map the feature vector set to the preset disturbance confidence map area, calculate the projection ratio and dynamic centroid offset in each early warning response domain, and determine whether the disturbance reaches the triggering standard corresponding to the perimeter alarm threshold of the detection cable. S5. Based on the judgment result, perform dynamic adjustment of the alarm threshold, and actively lower, gradually maintain, or keep the alarm threshold constant for different disturbance confidence levels.

2. The method for dynamically determining the perimeter alarm threshold of a detection cable based on dual-cable coupling according to claim 1, characterized in that, S1 specifically involves: after a disturbance occurs in the detection cable, simultaneously acquiring the original response data of the dual-cable coupling signal of the first and second cables, and performing amplitude normalization, time axis reconstruction, and data start point alignment processing on the response data based on a fixed time window to obtain a continuous signal segment for spectrum analysis; The raw response data of the dual-cable coupled signal is input into the time-frequency joint analysis process, and short-time Fourier transform and continuous wavelet transform are performed sequentially to extract the dominant frequency path, amplitude distribution and phase change information in multiple time periods, generating corresponding multi-time period spectrum sequences. In the multi-time period spectrum sequences, the positions where the rate of change of the dominant frequency trajectory on the frequency axis occurs are identified, and the frequency increment, amplitude difference and phase drift at the corresponding time points are extracted to construct a spectrum jump criterion set to reflect the starting point of coupling anomalies.

3. The method for dynamically determining the perimeter alarm threshold of a detection cable based on dual-cable coupling according to claim 1, characterized in that, S2 specifically includes the following steps: S201. Based on the spectrum jump criterion set, extract the main frequency scatter points corresponding to each time index, connect them in time order to form the main frequency scatter point trajectory of the dual-cable coupled signal, and calculate the offset distance and offset direction change of adjacent scatter points on the frequency axis respectively. By accumulating and smoothing the continuous offset changes, a consistency curvature is generated to describe the overall bending degree of the main frequency scatter point trajectory; S202. After completing the consistency curvature calculation, along the frequency interval corresponding to the main frequency scatter point trajectory, the number of times the same frequency component appears repeatedly in different time indices is counted to obtain the periodic repetition density, and the energy in the return direction and the energy in the forward propagation direction in the spectrum are integrated and accumulated to calculate the energy return ratio; S203. After obtaining the consistency curvature, periodic repetition density and energy return ratio, the three are compared with the pre-set structure judgment interval item by item. When the consistency curvature is in the high bending interval and the periodic repetition density and the energy return ratio both exceed the corresponding threshold, it is determined that the detection cable has crossed the electromagnetic reflection structure.

4. The method for dynamically determining the perimeter alarm threshold of a detection cable based on dual-cable coupling according to claim 3, characterized in that, S201 specifically involves: reading the position of the main frequency scatter points corresponding to each time index from the spectrum jump criterion set, arranging the main frequency scatter points in chronological order, and constructing a continuous trajectory of the main frequency scatter points of the dual-cable coupled signal based on adjacent time indices; The offset distance of adjacent main frequency scatter points on the frequency axis is calculated point by point along the main frequency scatter point trajectory, and the directional change between adjacent offset directions is calculated. The offset distance and directional change are combined to form a continuous offset change sequence. The continuous offset change sequence is processed by segmented accumulation, and a smoothing constraint is introduced during the accumulation process to suppress local discrete fluctuations. Based on the accumulation results, a uniform curvature is generated to characterize the overall bending degree of the main frequency scatter point trajectory.

5. The method for dynamically determining the perimeter alarm threshold of a detection cable based on dual-cable coupling according to claim 1, characterized in that, S3 specifically involves: when a detection cable passes through an electromagnetic reflection structure, extracting the response data of the signal channels corresponding to the first and second cables on the frequency axis from the dual-cable coupling signal; extracting the spectral amplitude trajectory and phase trajectory of the corresponding frequency band according to the time index alignment method to establish the frequency correspondence between the first and second cables under the same time index; comparing the dominant frequency position and phase center position of the signal channels of the first and second cables point by point along the frequency axis under each time index; calculating the frequency slip value and phase slip amount between each pair of response frequency points to form the relative slip distribution of the signal channels of the first and second cables on the frequency axis; based on the relative slip distribution of the signal channels of the first and second cables on the frequency axis, extracting the slip abrupt change point, slip direction inflection point, and slip duration interval; constructing a vector set containing the dominant frequency slip amount, slip interval length, slip point density, and phase slip amplitude; and generating a feature vector set to describe the spectral misalignment characteristics of the dual-cable coupling signal.

6. The method for dynamically determining the perimeter alarm threshold of a detection cable based on dual-cable coupling according to claim 1, characterized in that, S4 specifically includes the following steps: S401, standardize the feature vector set using a unified feature scale, and map the standardized feature vector set to a preset disturbance confidence map area. Within the preset disturbance confidence map area, divide multiple early warning response domains according to pre-defined response level boundaries, and assign the mapped coordinates corresponding to each feature vector to the corresponding early warning response domain; S402, statistically analyze the distribution of all feature vectors in each early warning response domain, calculate the ratio between the number of feature vectors in each early warning response domain and the total number of feature vectors, obtain the projection ratio in each early warning response domain, and calculate the weighted dynamic centroid position based on the coordinate distribution of all feature vectors in the disturbance confidence map area, and determine the degree of offset of the dynamic centroid relative to the reference center of the disturbance confidence map area; S403, compare the projection ratio and dynamic centroid offset in the highest early warning level response domain with the triggering standard corresponding to the perimeter alarm threshold of the detection cable, respectively. When the projection ratio in the highest early warning level response domain exceeds the corresponding ratio threshold and the dynamic centroid offset exceeds the corresponding offset threshold, it is determined that the disturbance has reached the triggering standard corresponding to the perimeter alarm threshold of the detection cable.

7. The method for dynamically determining the perimeter alarm threshold of a detection cable based on dual-cable coupling according to claim 6, characterized in that, S402 specifically involves: comparing the mapped coordinates of all feature vectors within the preset perturbation confidence map area with the boundary range of each early warning response domain; counting the number of feature vectors contained in each early warning response domain; and proportionally calculating this number with the total number of all feature vectors to obtain the projection ratio corresponding to each early warning response domain; based on the two-dimensional coordinate positions of all feature vectors within the preset perturbation confidence map area, constructing a weighted coordinate set with feature intensity as a weighting factor; calculating the weighted average of all weighted coordinates to obtain the weighted dynamic centroid position coordinates of the current perturbation event; selecting the coordinates of a reference center point set within the preset perturbation confidence map area; calculating the Euclidean distance between the weighted dynamic centroid position coordinates and the reference center point coordinates; and using this distance as the degree of dynamic centroid offset of the current perturbation event.

8. The method for dynamically determining the perimeter alarm threshold of a detection cable based on dual-cable coupling according to claim 1, characterized in that, S5 specifically involves: based on the judgment result, extracting the projection ratio and dynamic centroid shift degree of the feature vector set in the preset perturbation credibility map area, constructing a perturbation credibility level classification model based on the positional relationship between the two in the corresponding threshold, and classifying the current perturbation event into high perturbation credibility level, medium perturbation credibility level, or low perturbation credibility level. After obtaining the disturbance confidence level, dynamic adjustment of the alarm threshold is performed: when the disturbance confidence level is high, the alarm threshold is actively lowered; when the disturbance confidence level is medium, the alarm threshold is gradually maintained by using linear interpolation to smoothly adjust the alarm threshold within a preset adjustment period; when the disturbance confidence level is low, the alarm threshold is kept constant within the current period to prevent unnecessary alarms caused by low-confidence disturbances, thereby achieving dynamic adjustment of the alarm threshold under different disturbance confidence levels.

Images