Space electromagnetic wave feature extraction and identification method and device for lightning positioning system

Abstract

The invention discloses a space electromagnetic wave feature extraction and identification method and device for a lightning positioning system. The method comprises the following steps: acquiring a measured sampling sequence of measured lightning electromagnetic waves, and performing windowing, resampling to a target length L and normalization processing on the measured sampling sequence; constructing a reference prototype library; obtaining a sample sequence of each type of waveform from the reference prototype library; respectively determining similarity measurement between the tested sampling sequence and each sample sequence; outputting an identification result according to a comparison result of all the similarity measures and a preset threshold value; the identification result comprises the category label, the corresponding minimum similarity metric value and the optimal alignment translation amount.

Description

Technical Field

[0001] This invention relates to the field of circuit detection and signal analysis technology, and in particular to a method and apparatus for extracting and identifying spatial electromagnetic wave features for lightning location systems. Background Technology

[0002] The power industry is a vital component of the national economy, and its safe operation directly impacts the stability of people's lives and economic development. With the continuous expansion of the power grid, the threat of lightning strikes to the power system is becoming increasingly serious, especially in lightning-prone areas. Lines and equipment are highly susceptible to overvoltage or insulation flashover accidents caused by lightning strikes, leading to power outages and equipment damage, thus causing economic losses. Lightning electromagnetic wave signals, as important indicative information of lightning activity, can reflect the physical characteristics of different types of lightning processes. Their accurate identification is of great significance for the lightning protection design, operation and maintenance, and fault diagnosis of power systems.

[0003] Existing lightning monitoring mainly relies on lightning location systems, which use telemetry of the lightning electromagnetic field to locate the lightning strike point and calculate the lightning current amplitude. Although widely used, these systems still suffer from insufficient accuracy and robustness due to factors such as propagation path and environmental interference. Therefore, researching auxiliary identification methods based on electromagnetic wave waveforms is of great significance.

[0004] Currently, there are two main types of classification methods for lightning electromagnetic waves: one is based on manual parameter criteria, which makes judgments by extracting parameters such as waveform amplitude, pulse width, and rising edge, but it is prone to misjudgment when the feature intervals of different types of waveforms overlap; the other is based on deep learning intelligent classification methods, which can improve the accuracy of classification, but it is highly dependent on large-scale labeled data and high computing power platforms, making it difficult to deploy in low-computing-power online monitoring terminals.

[0005] Therefore, how to propose a spatial electromagnetic wave feature extraction and identification method for lightning location systems that can guarantee classification accuracy and robustness while reducing computational overhead and is suitable for deployment in online detection scenarios has become an urgent technical problem to be solved. Summary of the Invention

[0006] Therefore, it is necessary to propose a spatial electromagnetic wave feature extraction and identification method and device for lightning location systems to address the problems of complex lightning electromagnetic wave signals, susceptibility to noise, and difficulty in accurately distinguishing their categories in existing technologies.

[0007] A method for extracting and identifying spatial electromagnetic wave features for lightning location systems, the method comprising: The measured sampling sequence of the lightning electromagnetic wave is obtained, and the measured sampling sequence is subjected to windowing, resampling to the target length L, and normalization. A labeled sample library consisting of four types of waveforms, including ground flash return stroke (RS), cloud flash return stroke (IC), initial breakdown pulse (PB), and narrow bipolar pulse (NB), was constructed, as well as a reference prototype library containing a thousand sample sequences for each type of waveform. Obtain sample sequences of each type of waveform from the reference prototype library; Determine the similarity measure between the tested sampling sequence and each sample sequence; Based on the comparison results of all the similarity metrics with the preset threshold, the identification result is output; the identification result includes the category label, the corresponding minimum similarity metric value, and the optimal alignment translation amount.

[0008] Preferably, the step of performing windowing, resampling to the target length L, and normalization on the tested sample sequence specifically includes: Determine whether the sampling length of the tested sampling sequence exceeds the target length L; If the sampling length of the sampled sequence under test does not exceed the target length L, the sampled sequence under test with the original sampling length is used as input for normalization processing to obtain the sampled sequence under test. If the sampling length of the sampled sequence under test exceeds the target length L, the sampled sequence under test is resampled to shorten the sampling length to the target length L. Then, the sampled sequence under test with a sampling length of L is used as input for normalization processing to obtain the sampled sequence under test.

[0009] Preferably, the resampling of the tested sample sequence specifically includes: according to For length of N The sampled sequence under test Perform a discrete Fourier transform to obtain the spectrum. ,in, n=0,…,N-1,k=0,…,N 1 ; When the target length L >N At that time, the spectrum is analyzed while maintaining conjugate symmetry. Zero padding is performed to obtain a sampling length of L Spectrum ; When the target length L <N At that time, for the spectrum First multiply by the anti-aliasing low-pass window Then, while maintaining conjugate symmetry, symmetrical truncation is performed to obtain a sampling length of... L Spectrum ; according to For the sampling length isL Spectrum Perform a length of L The inverse discrete Fourier transform is used to obtain the resampled time-domain sequence. n=0,…,L 1; The resampled time-domain sequence The real part is the length of L The resampling results.

[0010] Preferably, the normalization process of the sampled sequence to obtain the sampled sequence specifically includes: according to Normalize each sampling point in the tested sampling sequence to obtain the mean of the tested sampling sequence. and standard deviation ;in, The first sampled sequence under test i sampling points Let S be the mean of all sampled points of the sequence being tested. The standard deviation of all sampling points in the tested sampling sequence is denoted as . The sampling points are normalized samples of the tested sampling sequence.

[0011] Preferably, determining the similarity measure between the tested sampling sequence and each sample sequence specifically includes: according to For each sampling point in the sample sequence, normalization is performed to obtain the mean of the sample sequence. and standard deviation ,in, For the sample sequence i One sampling point, The mean of all sample points in the sample sequence. The standard deviation of all sampling points in the sample sequence. These are the sample points after normalization in the sample sequence; according to Determine the normalized cross-correlation between the tested sampling sequence and the sample sequence. ,in, This is the translation amount; In all translations Below, the maximum value of the normalized cross-correlation is selected as the optimal alignment shift, and this optimal alignment shift is denoted as... , ; according to Determine the shape basis distance between the tested sampling sequence and each sample sequence, and obtain the minimum value among the shape basis distances. and second smallest value ;in, To determine the optimal alignment shift that maximizes the normalized cross-correlation, x Represents the sampled sequence being tested. r It represents a sample sequence.

[0012] Preferably, the step of outputting the identification result based on the comparison results of all the similarity measures with the preset threshold specifically includes: Determine the minimum value among the shape base distances. Is it greater than the preset threshold? i ; If the minimum value of the shape base distance Greater than the preset threshold i If so, the rejection result will be output directly; If the minimum value of the shape base distance Less than or equal to the preset threshold i Determine the minimum value among the shape base distances. and the second smallest value Is the difference between them greater than the preset interval? ; If the minimum value of the shape base distance and the second smallest value If the difference between them is greater than a preset interval, output the minimum value. The corresponding category label is used as the identification result; If the minimum value of the shape base distance and the second smallest value If the difference between the two values ​​is less than or equal to the preset interval, the rejection result will be output directly.

[0013] Preferably, after outputting the rejection result, the method further includes: Record the measured sampling sequence and the previous one P The shape base distance value SBD and the corresponding optimal translation amount are determined for each reference prototype. ; The recorded results are stored in a reference prototype library for subsequent dynamic updates of the execution threshold parameters and periodic reconstruction of the reference prototype library.

[0014] A spatial electromagnetic wave feature extraction and identification device for lightning location systems, the device comprising: The signal acquisition module is used to acquire the measured sampling sequence of the measured lightning electromagnetic wave as the input source of the device; The preprocessing module, connected to the signal acquisition module, is used to sequentially perform windowing, resampling to the target length L, and normalization processing on the sampled sequence under test. The storage module is used to store a reference prototype library, which includes a labeled sample library consisting of four types of waveforms: ground flash return stroke (RS), cloud flash return stroke (IC), initial breakdown pulse (PB), and narrow bipolar pulse (NB), as well as several sample sequences for each type of waveform. A similarity calculation module, connected to the preprocessing module and the storage module, is used to obtain sample sequences of each type of waveform from the reference prototype library and determine the similarity measure between the tested sampling sequence and each sample sequence. The determination output module is connected to the similarity calculation module and is used to output the identification result based on the comparison results of all the similarity measures with the preset threshold. The identification result includes the category label, the corresponding minimum similarity measure value, and the optimal alignment translation amount.

[0015] A computer device includes a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the following steps: The measured sampling sequence of the lightning electromagnetic wave is obtained, and the measured sampling sequence is subjected to windowing, resampling to the target length L, and normalization. A labeled sample library consisting of four types of waveforms, including ground flash return stroke (RS), cloud flash return stroke (IC), initial breakdown pulse (PB), and narrow bipolar pulse (NB), was constructed, as well as a reference prototype library containing a thousand sample sequences for each type of waveform. Obtain sample sequences of each type of waveform from the reference prototype library; Determine the similarity measure between the tested sampling sequence and each sample sequence; Based on the comparison results of all the similarity metrics with the preset threshold, the identification result is output; the identification result includes the category label, the corresponding minimum similarity metric value, and the optimal alignment translation amount.

[0016] A computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the following steps: The measured sampling sequence of the lightning electromagnetic wave is obtained, and the measured sampling sequence is subjected to windowing, resampling to the target length L, and normalization. A labeled sample library consisting of four types of waveforms, including ground flash return stroke (RS), cloud flash return stroke (IC), initial breakdown pulse (PB), and narrow bipolar pulse (NB), was constructed, as well as a reference prototype library containing a thousand sample sequences for each type of waveform. Obtain sample sequences of each type of waveform from the reference prototype library; Determine the similarity measure between the tested sampling sequence and each sample sequence; Based on the comparison results of all the similarity metrics with the preset threshold, the identification result is output; the identification result includes the category label, the corresponding minimum similarity metric value, and the optimal alignment translation amount.

[0017] Compared with existing technologies, this invention calculates the similarity between the tested sample sequence and sample sequences in a reference prototype library, and combines threshold and difference criteria to determine the category. This enables accurate identification of typical lightning electromagnetic wave waveforms such as ground flashback (RS), cloud flashback (IC), initial breakdown pulse (PB), and narrow bipolar pulse (NB) in actual monitoring environments. Unlike existing identification methods that rely on complex deep learning models or single threshold criteria, this invention achieves high-precision identification under low-computing-power conditions through unified preprocessing, similarity measurement, and reference prototype clustering mechanisms. This technical solution not only improves robustness to waveform translation and noise interference but also reduces dependence on large-scale training samples and high-performance computing platforms. Therefore, this invention can reduce system construction costs while ensuring identification reliability, making it suitable for widespread application in various engineering scenarios such as transmission line protection and online early warning systems, and has significant practical value and promotional significance. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] in: Figure 1 This is a flowchart illustrating a spatial electromagnetic wave feature extraction and identification method for a lightning location system, provided by an embodiment of the present invention. Figure 2 This invention provides a curve of the amplitude and difference between the sampled sequence and the sample sequence in a spatial electromagnetic wave feature extraction and identification method for lightning location systems, as provided in this embodiment of the invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] This invention provides a method for extracting and identifying spatial electromagnetic wave features for lightning location systems, such as... Figure 1 As shown, the method includes: Step 101: Obtain the measured sampling sequence of the measured lightning electromagnetic wave, and perform windowing, resampling to the target length L, and normalization processing on the measured sampling sequence. Specifically, it is determined whether the sampling length of the tested sampling sequence exceeds the target length L; If the sampling length of the sampled sequence under test does not exceed the target length L, the sampled sequence under test with the original sampling length is used as input for normalization processing to obtain the sampled sequence under test. If the sampling length of the sampled sequence under test exceeds the target length L, the sampled sequence under test is resampled to shorten the sampling length to the target length L. Then, the sampled sequence under test with a sampling length of L is used as input for normalization processing to obtain the sampled sequence under test.

[0022] according to For length of N The sampled sequence under test ( n=0,…,N-1 Perform a discrete Fourier transform to obtain the spectrum. ( k=0,…,N 1 ),in, ; When the target length L >N At that time, the spectrum is analyzed while maintaining conjugate symmetry. Zero padding is performed to obtain a sampling length of L Spectrum ; When the target length L <N At that time, for the spectrum First multiply by the anti-aliasing low-pass window Then, while maintaining conjugate symmetry, symmetrical truncation is performed to obtain a sampling length of... L Spectrum ; according to For the sampling length is L Spectrum Perform a length of L The inverse discrete Fourier transform is used to obtain the resampled time-domain sequence. n=0,…,L 1; The resampled time-domain sequence The real part is the length of L The resampling results.

[0023] Maintaining conjugate symmetry means satisfying: .

[0024] according to Normalize each sampling point in the tested sampling sequence to obtain the mean of the tested sampling sequence. and standard deviation ;in, The first sampled sequence under test i sampling points Let S be the mean of all sampled points of the sequence being tested. The standard deviation of all sampling points in the tested sampling sequence is denoted as . The sampling points are normalized samples of the tested sampling sequence.

[0025] Step 102: Construct a labeled sample library consisting of four types of waveforms: ground flash return stroke (RS), cloud flash return stroke (IC), initial breakdown pulse (PB), and narrow bipolar pulse (NB), as well as a reference prototype library containing a thousand sample sequences for each type of waveform. Specifically, the reference prototype library is generated through k-Shape clustering; The k-Shape clustering uses shape basis distance as a metric and generates cluster centroids through iterative centroid updates. These cluster centroids are then used as sample sequences for the corresponding categories. c The update definition is:

[0026]

[0027] in, For the first in the same sample set i sequence, In order to make The optimal alignment translation amount corresponding to the maximum value, the The optimal translation amount between the sample sequence and the candidate centroid. The shape base distance. The number of reference prototypes for each category is a positive integer. K , K The value range is 5 to 64, preferably 10 to 30, and the value of each type is... K Each reference prototype is a centroid waveform of a cluster obtained by clustering samples of this category.

[0028] Step 103: Obtain sample sequences of each type of waveform from the reference prototype library; Step 104: Determine the similarity measure between the tested sampling sequence and each sample sequence; Specifically, according to For each sampling point in the sample sequence, normalization is performed to obtain the mean of the sample sequence. and standard deviation ,in, For the sample sequence i One sampling point, The mean of all sample points in the sample sequence. The standard deviation of all sampling points in the sample sequence. These are the sample points after normalization in the sample sequence; according to Determine the normalized cross-correlation between the tested sampling sequence and the sample sequence. ,in, This is the translation amount; Among them, the tested sample sequence after z-score normalization and sample sequence Perform a Fast Fourier Transform to obtain the spectra respectively. and ,in, Indicates Fourier transform; Multiplying the two in the frequency domain and taking the complex conjugate spectrum of the sample sequence yields:

[0029] in, express The complex conjugate of the Fourier spectrum.

[0030] The resulting product Zero padding is used to ensure time-domain alignment accuracy; right Performing the inverse fast Fourier transform yields a linearly cross-correlated sequence:

[0031] in, Indicates the amount of translation. This represents the inverse fast Fourier transform.

[0032] The linear cross-correlation result is obtained by using the inverse fast Fourier transform, thus reducing the complexity of traditional time-domain computation. Reduced to Highly efficient computing.

[0033] In all translations Below, the maximum value of the normalized cross-correlation is selected as the optimal alignment shift, and this optimal alignment shift is denoted as... , ; according to Determine the shape basis distance between the tested sampling sequence and each sample sequence, and obtain the minimum value among the shape basis distances. and second smallest value ;in, To determine the optimal alignment shift that maximizes the normalized cross-correlation, x Represents the sampled sequence being tested. r It represents a sample sequence.

[0034] Step 105: Based on the comparison results of all the similarity metrics with the preset threshold, output the identification result; the identification result includes the category label, the corresponding minimum similarity metric value, and the optimal alignment translation amount.

[0035] Specifically, determine the minimum value among the shape base distances. Is it greater than the preset threshold? i ; If the minimum value of the shape base distance Greater than the preset threshold i If so, the rejection result will be output directly; If the minimum value of the shape base distance Less than or equal to the preset threshold i Determine the minimum value among the shape base distances. and the second smallest value Is the difference between them greater than the preset interval? ; If the minimum value of the shape base distance and the second smallest value The difference between them is greater than the preset interval Output the minimum value d The category label corresponding to 1 is used as the identification result; If the minimum value of the shape base distance and the second smallest value The difference between them is less than or equal to the preset interval. If the condition is not met, the rejection result will be output directly.

[0036] Furthermore, after outputting the rejection result, the method further includes: Record the measured sampling sequence and the previous one P The shape base distance value SBD and the corresponding optimal translation amount are determined for each reference prototype. ; The recorded results are stored in a reference prototype library for subsequent dynamic updates of the execution threshold parameters and periodic reconstruction of the reference prototype library.

[0037] This invention calculates the similarity between the tested rice planting sequence and sample sequences in a reference prototype library, and combines threshold and difference criteria to determine the category. This enables accurate identification of typical lightning electromagnetic wave waveforms such as ground flashback (RS), cloud flashback (IC), initial breakdown pulse (PB), and narrow bipolar pulse (NB) in actual monitoring environments. Unlike existing identification methods that rely on complex deep learning models or single threshold criteria, this invention achieves high-precision identification under low-computing-power conditions through unified preprocessing, similarity measurement, and reference prototype clustering mechanisms. This technical solution not only improves robustness to waveform translation and noise interference but also reduces dependence on large-scale training samples and high-performance computing platforms.

[0038] Therefore, this invention can reduce system construction costs while ensuring identification reliability, and is suitable for widespread application in various engineering scenarios such as power transmission line protection and online early warning systems, and has significant practical value and promotional significance.

[0039] Example 1: In an electromagnetic wave monitoring system deployed along a high-voltage transmission line corridor, electromagnetic wave signals from a lightning event were collected in the field, with a sampling rate set at 10 MHz. The measured waveform sequence obtained through the monitoring terminal device was used as input, and corresponding sample sequences were retrieved from a pre-built sample database. Subsequently, the measured data was processed and analyzed according to the method proposed in this invention, completing a typical spatial electromagnetic wave feature extraction and identification process for lightning location systems. The specific steps are as follows: (1) Perform windowing operation on the acquired lightning electromagnetic wave sequence and resample to the target length L = 5000 points, and perform normalization processing to obtain a standardized sequence.

[0040] (2) Retrieve similar sample sequences from the reference prototype library. The reference prototype library contains four typical lightning waveforms: RS, IC, PB, and NB. Each type consists of multiple prototype sequences generated by aggregation within the type.

[0041] (3) Calculate the similarity measure between the tested sequence and each sample sequence, obtain the category label corresponding to the maximum similarity, and extract the translation amount under the optimal match. In all similarity calculations, the Fast Fourier Transform is used for acceleration.

[0042] (4) Set the threshold θ = 0.2 and the interval Δ = 0.05. If the distance corresponding to the maximum similarity is less than the threshold and the difference between the maximum similarity and the second minimum similarity is greater than the interval, then output the matching result; otherwise, output rejection.

[0043] (5) In this embodiment, the tested sequence has the highest similarity to the IC class reference prototype. The minimum distance value is 0.12, which is less than the threshold θ = 0.2, and the difference between it and the second smallest value is 0.08, which is greater than the interval Δ = 0.05. Therefore, the output category is IC.

[0044] Furthermore, regarding the complexity of similarity calculation, if cross-correlation is calculated directly in the time domain, it requires approximately 5.0 × 10⁻⁶ samples with L = 5000. 7 After acceleration using FFT, this operation requires only approximately 3.46 × 10^6 operations. 6 The complexity of the complex operation is... Down to This achieved an acceleration of approximately 14 times.

[0045] like Figure 2 As shown, this invention adopts a modular system structure, decoupling the acquisition, preprocessing, feature calculation, and classification output of lightning electromagnetic wave signals, thus achieving a layered design of the algorithm and platform. This structure allows the invention to be deployed on edge computing terminals and electromagnetic monitoring station field equipment, reducing computing resource consumption while ensuring real-time performance, and improving system scalability and deployment capabilities.

[0046] Conclusion: Through simulation and calculation in this embodiment, it is demonstrated that the present invention can accurately identify typical lightning electromagnetic wave waveforms under limited sample conditions and low computing power. Compared with traditional methods that rely on complex neural networks, the present invention has the advantages of simple implementation, low computational overhead, and strong robustness to waveform translation, making it suitable for online real-time monitoring and edge computing scenarios.

[0047] Example 2: In an electromagnetic wave monitoring system deployed along a high-voltage transmission line corridor, electromagnetic wave signals from a lightning event were collected in the field, with a sampling rate set at 10 MHz. The measured waveform sequence obtained through the monitoring terminal device was used as input, and corresponding sample sequences were retrieved from a pre-built sample database. Subsequently, the measured data was processed and analyzed according to the method proposed in this invention, completing a typical spatial electromagnetic wave feature extraction and identification process for lightning location systems. The specific steps are as follows: (1) Perform windowing operation on the acquired lightning electromagnetic wave sequence and resample to the target length L = 5000 points, and perform normalization processing to obtain a standardized sequence.

[0048] (2) Retrieve similar sample sequences from the reference prototype library. The reference prototype library contains four typical lightning waveforms: RS, IC, PB, and NB. Each type consists of multiple prototype sequences generated by aggregation within the type.

[0049] (3) Calculate the similarity measure between the tested sequence and each sample sequence, obtain the category label corresponding to the maximum similarity, and extract the translation amount under the optimal match.

[0050] (4) Set the threshold θ = 0.20 and the interval Δ = 0.05. If the distance corresponding to the maximum similarity is less than the threshold and the difference between the maximum similarity and the second minimum similarity is greater than the interval, then output the matching result; otherwise, output rejection.

[0051] (5) In this embodiment, the tested sequence has the highest similarity to the IC class reference prototype. The minimum distance value is 0.17, which is less than the threshold θ = 0.2, and the difference between it and the second smallest value is 0.07, which is greater than the interval Δ = 0.05. Therefore, the output category is IC.

[0052] Furthermore, regarding the complexity of similarity calculation, if cross-correlation is calculated directly in the time domain, it requires approximately 5.0 × 10⁻⁶ samples with L = 5000. 7 After acceleration using FFT, this operation requires only approximately 3.46 × 10^6 operations. 6 The complexity of the complex operation is... Down to This achieved an acceleration of approximately 14 times.

[0053] Conclusion: Through simulation and calculation in this embodiment, it is demonstrated that the present invention can accurately identify typical lightning electromagnetic wave waveforms under limited sample conditions and low computing power. Compared with traditional methods that rely on complex neural networks, the present invention has the advantages of simple implementation, low computational overhead, and strong robustness to waveform translation, making it suitable for online real-time monitoring and edge computing scenarios.

[0054] This invention also provides a spatial electromagnetic wave feature extraction and identification device for a lightning location system, the device comprising: The signal acquisition module is used to acquire the measured sampling sequence of the measured lightning electromagnetic wave as the input source of the device; The preprocessing module, connected to the signal acquisition module, is used to sequentially perform windowing, resampling to the target length L, and normalization processing on the sampled sequence under test. The storage module is used to store a reference prototype library, which includes a labeled sample library consisting of four types of waveforms: ground flash return stroke (RS), cloud flash return stroke (IC), initial breakdown pulse (PB), and narrow bipolar pulse (NB), as well as several sample sequences for each type of waveform. A similarity calculation module, connected to the preprocessing module and the storage module, is used to obtain sample sequences of each type of waveform from the reference prototype library and determine the similarity measure between the tested sampling sequence and each sample sequence. The determination output module is connected to the similarity calculation module and is used to output the identification result based on the comparison results of all the similarity measures with the preset threshold. The identification result includes the category label, the corresponding minimum similarity measure value, and the optimal alignment translation amount.

[0055] This invention also provides a computer device, including a memory and a processor. The memory stores a computer program, and when the computer program is executed by the processor, the processor performs the following steps: Step 101: Obtain the measured sampling sequence of the measured lightning electromagnetic wave, and perform windowing, resampling to the target length L, and normalization processing on the measured sampling sequence. Step 102: Construct a labeled sample library consisting of four types of waveforms: ground flash return stroke (RS), cloud flash return stroke (IC), initial breakdown pulse (PB), and narrow bipolar pulse (NB), as well as a reference prototype library containing a thousand sample sequences for each type of waveform. Step 103: Obtain sample sequences of each type of waveform from the reference prototype library; Step 104: Determine the similarity measure between the tested sampling sequence and each sample sequence; Step 105: Based on the comparison results of all the similarity metrics with the preset threshold, output the identification result; the identification result includes the category label, the corresponding minimum similarity metric value, and the optimal alignment translation amount.

[0056] This invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the following steps: Step 101: Obtain the measured sampling sequence of the measured lightning electromagnetic wave, and perform windowing, resampling to the target length L, and normalization processing on the measured sampling sequence. Step 102: Construct a labeled sample library consisting of four types of waveforms: ground flash return stroke (RS), cloud flash return stroke (IC), initial breakdown pulse (PB), and narrow bipolar pulse (NB), as well as a reference prototype library containing a thousand sample sequences for each type of waveform. Step 103: Obtain sample sequences of each type of waveform from the reference prototype library; Step 104: Determine the similarity measure between the tested sampling sequence and each sample sequence; Step 105: Based on the comparison results of all the similarity metrics with the preset threshold, output the identification result; the identification result includes the category label, the corresponding minimum similarity metric value, and the optimal alignment translation amount.

[0057] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), and double data rate RAM. The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification. The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for extracting and identifying spatial electromagnetic wave features for lightning location systems, comprising an acquisition module, a preprocessing module, and a reference prototype library, characterized in that: The measured sampling sequence of the lightning electromagnetic wave is obtained, and the measured sampling sequence is subjected to windowing, resampling to the target length L, and normalization. A labeled sample library consisting of four types of waveforms, including ground flash return stroke (RS), cloud flash return stroke (IC), initial breakdown pulse (PB), and narrow bipolar pulse (NB), was constructed, as well as a reference prototype library containing a thousand sample sequences for each type of waveform. Obtain sample sequences of each type of waveform from the reference prototype library; Determine the similarity measure between the tested sampling sequence and each sample sequence; Based on the comparison results of all the aforementioned similarity measures with the preset threshold, the identification result is output; The identification results include the category label, the corresponding minimum similarity metric, and the optimal alignment translation amount.

2. The spatial electromagnetic wave feature extraction and identification method for lightning location systems according to claim 1, wherein the step of performing windowing, resampling to the target length L, and normalization processing on the sampled sequence under test is characterized in that: Determine whether the sampling length of the tested sampling sequence exceeds the target length L; If the sampling length of the sampled sequence under test does not exceed the target length L, the sampled sequence under test with the original sampling length is used as input for normalization processing to obtain the sampled sequence under test. If the sampling length of the sampled sequence under test exceeds the target length L, the sampled sequence under test is resampled to shorten the sampling length to the target length L. Then, the sampled sequence under test with a sampling length of L is used as input for normalization processing to obtain the sampled sequence under test.

3. The spatial electromagnetic wave feature extraction and identification method for lightning location systems according to claim 2, wherein resampling of the measured sampling sequence is performed, characterized in that: according to For length of N The sampled sequence under test Perform a discrete Fourier transform to obtain the spectrum. ,in, n=0,…,N-1, k=0,…,N 1 ; When the target length L >N At that time, the spectrum is analyzed while maintaining conjugate symmetry. Zero padding is performed to obtain a sampling length of L Spectrum ; When the target length L <N At that time, for the spectrum First multiply by the anti-aliasing low-pass window Then, while maintaining conjugate symmetry, symmetrical truncation is performed to obtain a sampling length of... L Spectrum ; according to For the sampling length is L Spectrum Perform a length of L The inverse discrete Fourier transform is used to obtain the resampled time-domain sequence. n=0,…,L 1; The resampled time-domain sequence The real part is the length of L The resampling results.

4. The spatial electromagnetic wave feature extraction and identification method for lightning location systems according to any one of claims 1-3, wherein the normalization processing of the sampled sequence to obtain the sampled sequence is characterized in that: according to Normalize each sampling point in the tested sampling sequence to obtain the mean of the tested sampling sequence. and standard deviation ;in, The first sampled sequence under test i sampling points Let S be the mean of all sampled points of the sequence being tested. The standard deviation of all sampling points in the tested sampling sequence is denoted as . The sampling points are normalized samples of the tested sampling sequence.

5. The spatial electromagnetic wave feature extraction and identification method for lightning location systems according to claim 1, wherein the similarity metric between the measured sampling sequence and each sample sequence is determined, characterized in that: according to For each sampling point in the sample sequence, normalization is performed to obtain the mean of the sample sequence. and standard deviation ,in, For the sample sequence i One sampling point, The mean of all sample points in the sample sequence. The standard deviation of all sampling points in the sample sequence. These are the sample points after normalization in the sample sequence; according to Determine the normalized cross-correlation between the tested sampling sequence and the sample sequence. ,in, This is the translation amount; In all translations Below, the maximum value of the normalized cross-correlation is selected as the optimal alignment shift, and this optimal alignment shift is denoted as... , ; according to Determine the shape basis distance between the tested sampling sequence and each sample sequence, and obtain the minimum value among the shape basis distances. and second smallest value ;in, To determine the optimal alignment shift that maximizes the normalized cross-correlation, x Represents the sampled sequence being tested. r It represents a sample sequence.

6. The spatial electromagnetic wave feature extraction and identification method for lightning location systems according to claim 5, wherein the identification result is output based on the comparison result of similarity measurement and preset threshold, characterized in that: Determine the minimum value among the shape base distances. Is it greater than the preset threshold? θ ; If the minimum value of the shape base distance Greater than the preset threshold θ If so, the rejection result will be output directly; If the minimum value of the shape base distance Less than or equal to the preset threshold θ Determine the minimum value among the shape base distances. and the second smallest value Is the difference between them greater than the preset interval? If the minimum value of the shape base distance and the second smallest value If the difference between them is greater than a preset interval, output the minimum value. The corresponding category label is used as the identification result; If the minimum value of the shape base distance and the second smallest value The difference between them is less than or equal to the preset interval. If the condition is not met, the rejection result will be output directly.

7. The spatial electromagnetic wave feature extraction and identification method for lightning location systems according to claim 6, wherein after outputting the rejection result, the method is characterized in that: Record the measured sampling sequence and the previous one P The shape base distance value SBD and the corresponding optimal translation amount are determined for each reference prototype. ; The recorded results are stored in a reference prototype library for subsequent dynamic updates of the execution threshold parameters and periodic reconstruction of the reference prototype library.

8. A spatial electromagnetic wave feature extraction and identification device for lightning location systems, characterized in that, The device includes an acquisition module, a preprocessing module, a reference prototype library, a similarity calculation module, and a decision output module, wherein: The signal acquisition module is used to acquire the measured sampling sequence of the measured lightning electromagnetic wave as the input source of the device; The preprocessing module, connected to the signal acquisition module, is used to sequentially perform windowing, resampling to the target length L, and normalization processing on the sampled sequence under test. The storage module is used to store a reference prototype library, which includes a labeled sample library consisting of four types of waveforms: ground flash return stroke (RS), cloud flash return stroke (IC), initial breakdown pulse (PB), and narrow bipolar pulse (NB), as well as several sample sequences for each type of waveform. A similarity calculation module, connected to the preprocessing module and the storage module, is used to obtain sample sequences of each type of waveform from the reference prototype library and determine the similarity measure between the tested sampling sequence and each sample sequence. The determination output module is connected to the similarity calculation module and is used to output the identification result based on the comparison results of all the similarity measures with the preset threshold. The identification result includes the category label, the corresponding minimum similarity measure value, and the optimal alignment translation amount.

9. A computer device for extracting and identifying spatial electromagnetic wave features in a lightning location system, characterized in that, The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor, when executing the computer program, implements the spatial electromagnetic wave feature extraction and identification method for a lightning location system as described in any one of claims 1 to 7, to perform windowing, resampling, and normalization processing on the sampled sequence of the measured lightning electromagnetic wave, and outputs the identification result.

10. A computer-readable storage medium for spatial electromagnetic wave feature extraction and identification in lightning location systems, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the lightning electromagnetic wave waveform identification method as described in any one of claims 1 to 7, so as to extract and classify the features of the measured lightning electromagnetic wave signal through program execution.

Images