Space electromagnetic wave identification method and device for intelligent lightning positioning system
By performing noise reduction and Fourier transform on lightning electromagnetic waves, combined with L2 normalization and similarity fusion, the problem of misjudgment in the classification of lightning radiated electric fields is solved, and high-precision and low-complexity lightning waveform recognition is achieved.
Patent Information
- Application Number
- CN202511693618.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-18
- Publication Date
- 2026-02-13
AI Technical Summary
Traditional lightning radiation electric field classification methods suffer from misclassification in fine classification tasks, and machine learning models are complex and energy-intensive when deployed on edge devices, making it difficult to achieve high accuracy and real-time performance.
By denoising and windowing the VLF-MF lightning electromagnetic wave, performing Fast Fourier Transform (FFT) and L2 normalization, calculating the FFT-cosine segment similarity, and then using a fusion module to weightedly fuse multi-subband information, iterative identification is performed using a preset threshold.
It achieves stable differentiation of lightning electromagnetic waves in complex environments, improves the accuracy and real-time performance of lightning classification, and reduces computational complexity and recognition time.
Smart Images
Figure CN121522271A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application belongs to the field of lightning monitoring and protection of power systems, and particularly relates to a spatial electromagnetic wave identification method and device for an intelligent lightning positioning system. BACKGROUND
[0002] Lightning is a typical atmospheric discharge phenomenon, and the electromagnetic radiation generated thereby covers a very wide frequency band. The signals in the VLF-MF frequency band are widely used in the field of lightning monitoring and protection of power systems because of their long propagation distance, small attenuation and easy reception.
[0003] Accurate classification of lightning radiation electric fields of different lightning activities not only provides a solid foundation for lightning protection design, but also provides data support for in-depth study of the development process of lightning. However, the traditional criterion method relying on a small number of time domain parameter thresholds has systematic limitations in fine classification tasks. PB, NBE and other cloud flash processes are distributed in overlapping ranges in terms of key parameters such as amplitude and pulse width, resulting in a large number of misjudgments of "cloud flash local flash". Existing statistics show that about 18% of RS in NLDN data are actually PB, and >97% of NBE are misjudged as RS. At the same time, lightning waveform parameters show significant regional differences, making it difficult to form stable and universal threshold boundaries. To reduce the bias of artificial feature selection and improve the generalization performance, some scholars have introduced machine learning and deep learning and achieved high accuracy, but such models usually have large parameter quantities and high computing power consumption, and their deployment in edge devices and real-time scenarios still faces complexity and energy consumption constraints.
[0004] To further improve the multi-classification accuracy of lightning radiation electric fields, the present application proposes a method with low computational complexity and easy engineering deployment, which aims to maintain the ability to distinguish four types of lightning electromagnetic waves in complex environments, and can be smoothly connected with existing wide-area monitoring and positioning systems to improve the overall reliability and real-time performance of classification, positioning and current inversion. SUMMARY
[0005] Therefore, it is necessary to propose a spatial electromagnetic wave identification method and device for an intelligent lightning positioning system in view of the above problems.
[0006] A spatial electromagnetic wave identification method for an intelligent lightning positioning system, the method comprising: Obtaining VLF-MF lightning electromagnetic waves to be identified, and performing noise reduction and window function processing on the waveforms of the VLF-MF lightning electromagnetic waves; Performing fast Fourier transform on the VLF-MF lightning electromagnetic waves to obtain the frequency spectrum features of the VLF-MF lightning electromagnetic waves; Performing L2 normalization processing on the frequency spectrum features of the VLF-MF lightning electromagnetic waves; A lightning waveform library containing four typical VLF-MF lightning electromagnetic waves is pre-constructed, including Return Stroke (RS), Preliminary Breakdown (PB), Narrow Bipolar Event (NBE), and Intracloud Lightning (IC). The FFT-cosine segment similarity is determined by comparing the normalized spectral characteristics with a pre-built lightning waveform library one by one. The FFT-cosine segment similarity is obtained by fusing all FFT-cosine segment similarities. Based on the FFT-cosine segmentation fusion similarity and a preset threshold Matching results of The lightning electromagnetic wave type of the VLF-MF lightning electromagnetic wave is determined.
[0007] Preferably, the L2 normalization processing of the spectral characteristics of the VLF-MF lightning electromagnetic wave specifically includes: according to , Determine the L2 norm of the spectral characteristics of the VLF-MF lightning electromagnetic wave respectively; L2 normalization is performed based on the L2 norm of the spectral characteristics of the VLF-MF lightning electromagnetic wave.
[0008] Preferably, the step of determining the FFT-cosine segment similarity by comparing the normalized spectral features with a pre-built lightning waveform library one by one specifically includes: according to Determine FFT-cosine segment similarity ;in, X The spectral vector of the VLF-MF lightning wave is... Y i The spectral vector of any VLF–MF lightning wave from a pre-built lightning waveform library.
[0009] Preferably, the step of fusing all FFT-cosine segment similarities to obtain FFT-cosine segment fusion similarity specifically includes: according to Obtain FFT-cosine segmented fusion similarity ;in, The weighted value for the FFT-cosine segment fusion similarity. .
[0010] Preferably, the sampling rate of the VLF-MF lightning electromagnetic wave to be identified is consistent with the sampling rate of the VLF–MF lightning wave in the lightning waveform library.
[0011] Preferably, the lightning electromagnetic wave type of the VLF-MF lightning electromagnetic wave is determined according to the matching result of the FFT-cosine subsection fusion similarity and the preset threshold, and specifically includes: comparing the FFT-cosine subsection fusion similarity S i with the preset threshold S a ; If S i is less than S a , the current VLF-MF lightning wave in the lightning waveform library is not matched, and the similarity with the next VLF-MF lightning wave in the lightning waveform library is continued to be calculated; If S i is greater than or equal to S a , the current VLF-MF lightning wave in the lightning waveform library is matched, and the lightning electromagnetic wave type of the current VLF-MF lightning wave is taken as the recognition result.
[0012] Preferably, the preset threshold S a is determined according to historical samples or test set experimental results.
[0013] A spatial electromagnetic wave recognition device for an intelligent lightning positioning system, the device comprising: a data acquisition and preprocessing module, configured to acquire a VLF-MF lightning electromagnetic wave to be recognized, and to perform noise reduction and window function processing on a waveform of the VLF-MF lightning electromagnetic wave; to perform fast Fourier transform on the VLF-MF lightning electromagnetic wave to obtain a frequency spectrum feature of the VLF-MF lightning electromagnetic wave; and to perform L2 normalization processing on the frequency spectrum feature of the VLF-MF lightning electromagnetic wave; a lightning waveform library construction module, configured to pre-construct a lightning waveform library containing four typical VLF-MF lightning electromagnetic waves of ground flash return stroke (RS), preliminary breakdown (PB), narrow bipolar event (NBE), and intracloud lightning (IC); a similarity calculation module, configured to compare the normalized frequency spectrum feature with the pre-constructed lightning waveform library one by one to determine an FFT-cosine subsection similarity; a fusion module configured to fuse all the FFT-cosine segment similarities to obtain an FFT-cosine segment fusion similarity; a matching module configured to determine the lightning electromagnetic wave type of the VLF-MF lightning electromagnetic wave according to a matching result of the FFT-cosine segment fusion similarity and a preset threshold.
[0014] A computer device comprising a memory and a processor, the memory storing a computer program, the computer program being executed by the processor to cause the processor to perform the following steps: obtaining a VLF-MF lightning electromagnetic wave to be identified, and performing noise reduction and window function processing on a waveform of the VLF-MF lightning electromagnetic wave; performing fast Fourier transform on the VLF-MF lightning electromagnetic wave to obtain a frequency spectrum feature of the VLF-MF lightning electromagnetic wave; performing L2 normalization processing on the frequency spectrum feature of the VLF-MF lightning electromagnetic wave; pre-constructing a lightning waveform library containing four types of typical VLF-MF lightning electromagnetic waves, i.e., ground flash return stroke (RS), preliminary breakdown (PB), narrow bipolar event (NBE), and intracloud lightning (IC); comparing the normalized frequency spectrum feature with the pre-constructed lightning waveform library one by one to determine an FFT-cosine segment similarity; fusing all the FFT-cosine segment similarities to obtain an FFT-cosine segment fusion similarity; determining the lightning electromagnetic wave type of the VLF-MF lightning electromagnetic wave according to a matching result of the FFT-cosine segment fusion similarity and a preset threshold. Matching results of
[0015] A computer readable storage medium storing a computer program, the computer program being executed by a processor to cause the processor to perform the following steps: obtaining a VLF-MF lightning electromagnetic wave to be identified, and performing noise reduction and window function processing on a waveform of the VLF-MF lightning electromagnetic wave; performing fast Fourier transform on the VLF-MF lightning electromagnetic wave to obtain a frequency spectrum feature of the VLF-MF lightning electromagnetic wave; performing L2 normalization processing on the frequency spectrum feature of the VLF-MF lightning electromagnetic wave; A lightning waveform library is constructed in advance, which contains four types of typical VLF-MF lightning electromagnetic waves, i.e., Return Stroke (RS), Preliminary Breakdown (PB), Narrow Bipolar Event (NBE) and Intracloud Lightning (IC); The FFT-cosine subsection similarity is determined by comparing the normalized spectrum feature with the lightning waveform library one by one; The FFT-cosine subsection fusion similarity is obtained by fusing all the FFT-cosine subsection similarities; The lightning electromagnetic wave type of the VLF-MF lightning electromagnetic wave is determined according to the FFT-cosine subsection fusion similarity and a preset threshold value Matching results of
[0016] Compared with the prior art, the VLF-MF lightning electromagnetic wave is denoised and window function processed, noise and spectrum leakage are effectively suppressed, clean input is provided for feature extraction, then the spectrum feature is obtained through fast Fourier transform, the amplitude difference is eliminated through L2 normalization processing of the data standardization module, the recognition only depends on the spectrum shape, the robustness of the method in a complex environment is improved, then the FFT-cosine subsection similarity is calculated through comparison with the lightning waveform library, the multi-subband information is weighted and fused through the fusion module, the discrimination accuracy is enhanced, finally, the iterative fast recognition is realized through threshold comparison, if the similarity is insufficient, the comparison is automatically continued, the time consumption of single sample recognition is less than 1 ms while the high precision is maintained. BRIEF DESCRIPTION OF DRAWINGS
[0017] Figure 1 A flowchart of a spatial electromagnetic wave recognition method for an intelligent lightning positioning system is provided for the embodiment of the present application; Figure 2 A time sequence waveform diagram of a VLF-MF lightning electromagnetic wave in the spatial electromagnetic wave recognition method for the intelligent lightning positioning system is provided for the embodiment 1 of the present application; Figure 3 A waveform diagram of a VLF-MF lightning electromagnetic wave after preprocessing in the spatial electromagnetic wave recognition method for the intelligent lightning positioning system is provided for the embodiment 1 of the present application; Figure 4 A spectrum feature diagram of a VLF-MF lightning electromagnetic wave in the spatial electromagnetic wave recognition method for the intelligent lightning positioning system is provided for the embodiment 1 of the present application; Figure 5 A time sequence waveform chart of a VLF-MF lightning electromagnetic wave in a spatial electromagnetic wave recognition method for an intelligent lightning positioning system is provided for Embodiment 2 of the present application; Figure 6 A preprocessed waveform chart of a VLF-MF lightning electromagnetic wave in a spatial electromagnetic wave recognition method for an intelligent lightning positioning system is provided for Embodiment 2 of the present application; Figure 7 A VLF-MF lightning electromagnetic wave spectrum feature chart in a spatial electromagnetic wave recognition method for an intelligent lightning positioning system is provided for Embodiment 2 of the present application. DETAILED DESCRIPTION
[0018] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.
[0019] Embodiments of the present application provide a spatial electromagnetic wave recognition method for an intelligent lightning positioning system, as shown in Figure 1 The method comprises the following steps: Step 101: acquiring a VLF-MF lightning electromagnetic wave to be recognized, and performing noise reduction and window function processing on a waveform of the VLF-MF lightning electromagnetic wave; Specifically, a sampling rate of the VLF-MF lightning electromagnetic wave to be recognized is consistent with a sampling rate of a VLF-MF lightning waveform in a lightning waveform library.
[0020] Step 102: performing fast Fourier transform on the VLF-MF lightning electromagnetic wave to obtain a spectrum feature of the VLF-MF lightning electromagnetic wave; Step 103: performing L2 normalization processing on the spectrum feature of the VLF-MF lightning electromagnetic wave; Specifically, L2 norms of the spectrum feature of the VLF-MF lightning electromagnetic wave are determined according to 、 respectively. L2 normalization processing is performed according to the L2 norms of the spectrum feature of the VLF-MF lightning electromagnetic wave.
[0021] Step 104, pre-constructing a lightning waveform library containing four types of typical VLF-MF lightning electromagnetic waves, including ground flash return stroke (Return Stroke, RS), preliminary breakdown pulse (Preliminary Breakdown, PB), narrow bipolar event (Narrow Bipolar Event, NBE), and intracloud lightning (Intracloud Lightning, IC); Step 105, comparing the normalized spectrum features with the pre-constructed lightning waveform library one by one to determine the FFT-cosine segment similarity; Specifically, the FFT-cosine segment similarity is obtained by dividing the frequency band into multiple sub-bands and calculating the sub-band similarity respectively.
[0022] According to determining the FFT-cosine segment similarity ; wherein, X is the spectrum vector of the VLF-MF lightning wave, Y i is the spectrum vector of any VLF-MF lightning wave in the pre-constructed lightning waveform library.
[0023] The spectrum vector of the VLF-MF lightning wave selects the frequency range of 3 kHz-3 MHz.
[0024] Step 106, fusing all FFT-cosine segment similarities to obtain FFT-cosine segment fusion similarity; Specifically, according to obtaining the FFT-cosine segment fusion similarity ; wherein, is the weighted value of the FFT-cosine segment fusion similarity, .
[0025] The weighted value The weight selection is based on the propagation stability, and higher weight is set for low-frequency sub-band, and lower weight is given to high-frequency attenuation serious sub-band.
[0026] Step 107, determining the lightning electromagnetic wave type of the VLF-MF lightning electromagnetic wave according to the FFT-cosine segment fusion similarity and a preset threshold Matching results of .
[0027] Specifically, comparing the size of the FFT-cosine segment fusion similarity S i with the preset threshold S a . If S i Less than S a If not matched with the current VLF-MF lightning waveform in the lightning waveform library, continue to calculate the similarity with the next VLF-MF lightning waveform in the lightning waveform library; If S i Greater than or equal to S a If matched with the current VLF-MF lightning waveform in the lightning waveform library, take the lightning electromagnetic wave type of the current VLF-MF lightning waveform as the identification result.
[0028] The preset threshold S a According to historical samples or test set experimental results, the range is 0.85-0.95.
[0029] The application can fully utilize the spectral characteristics of lightning electromagnetic waves in the very low frequency to medium frequency band, realize rapid, accurate and robust identification of typical lightning waveforms, and is suitable for real-time large-scale lightning monitoring.
[0030] The application uses the sampling FFT-cosine subsection fusion similarity algorithm for calculation, the overall complexity of the algorithm is low, the identification time of a single sample is less than 1ms, and the identification speed is fast; The application performs L2 normalization on the spectral feature vector, effectively eliminates the amplitude difference caused by different detection stations and different propagation paths, makes the identification only rely on the spectral energy distribution shape, and thus guarantees the stability and universality of the method in a complex electromagnetic propagation environment; The application uses the main energy distribution characteristics of the VLF-MF frequency band, combines single-band and sub-band multi-level similarity calculation and weighted fusion strategy, can accurately distinguish typical lightning electromagnetic wave types such as RS, PB, NBE and IC, and test set data verification shows that the method has high accuracy.
[0031] Embodiment 1: The pre-constructed lightning waveform library contains four typical VLF-MF lightning electromagnetic waves, namely, ground flash return stroke (Return Stroke, RS), preliminary breakdown pulse (Preliminary Breakdown, PB), narrow bipolar pulse (Narrow Bipolar Event, NBE) and intracloud lightning (Intracloud Lightning, IC). Each type of waveform contains 4000 data, the data sampling rate is 10 MHz, the time resolution is 0.1 μs, the sampling time window is 2 ms, the sampling point number is 20000, and the sampling time interval is 0.1 μs.
[0032] The specific steps of identification are as follows: 1. Obtain VLF-MF lightning electromagnetic waves with a sampling rate of 10 MHz and a sampling time window of 2 ms through a lightning monitoring system, and the time sequence waveform of the lightning waves is as shown in Figure 2 . 2. The lightning electromagnetic waves are subjected to wavelet denoising, and are subjected to Hann Window processing to suppress spectral leakage and are subjected to normalization processing, and the processed VLF-MF lightning electromagnetic wave waveform is as shown in Figure 3 .
[0033] 3. The pre-processed VLF-MF lightning electromagnetic waves are subjected to fast Fourier transform to obtain a single-side amplitude spectrum of 3 kHz-3 MHz.
[0034] 4. The amplitude spectrum vector is subjected to L2 normalization processing to eliminate the influence of amplitude difference, and the VLF-MF lightning electromagnetic wave spectrum feature is as shown in Figure 4 .
[0035] 5. The VLF-MF lightning waveform is compared with the pre-constructed four types of typical VLF-MF lightning waveform library one by one, and the FFT-cosine subsection similarity is calculated S b .
[0036] The 3 kHz-3 MHz frequency band is divided into three sub-bands: a low-frequency sub-band of 3 kHz-30 kHz, a medium-frequency sub-band of 30 kHz-300 kHz, and a high-frequency sub-band of 300 kHz-3 MHz, and the FFT-cosine subsection similarity of the three sub-bands is calculated by formula (1). The calculation results of the matching waveforms are as follows: 1) Low-frequency sub-band FFT-cosine subsection similarity S1=0.944 2) Medium-frequency sub-band FFT-cosine subsection similarity S2=0.806 3) Medium-frequency sub-band FFT-cosine subsection similarity S3=0.974 6. Calculate the FFT-cosine subsection fusion similarity S i .
[0037] The weighting value of the FFT-cosine subsection fusion similarity K b The weight selection is based on the propagation stability, a higher weight is set for the low-frequency sub-band, and a lower weight is given to the high-frequency attenuation serious sub-band, so the low-frequency sub-band weighting value K 1 is 0.7, the medium-frequency sub-band weighting value K 2 is 0.2, and the high-frequency sub-band weighting value K3 is set to 0.1. The calculated FFT-cosine segment similarity is then used. S b Weighted value of similarity with FFT-cosine segment fusion K b Substitute into formula (4) to calculate the FFT-cosine segment fusion similarity. S i The calculation results of the matched waveform are as follows: (9) The FFT-cosine segmented fusion similarity S is determined to match a preset threshold Sa, thereby quickly verifying the VLF-MF lightning electromagnetic wave type. The preset threshold for identification... S a Based on historical samples and test set experimental results, a value of 0.85 was selected, and the calculated waveform matching result was obtained. S i Greater than S a The identification result is: the waveform is a return stroke (RS) electromagnetic wave.
[0038] Example 2: The pre-built lightning waveform library contains four typical VLF-MF lightning electromagnetic waves: Return Stroke (RS), Preliminary Breakdown (PB), Narrow Bipolar Event (NBE), and Intracloud Lightning (IC). Each waveform contains 4000 data points, with a data sampling rate of 10 MHz, a time resolution of 0.1 μs, a sampling window of 2 ms, 20000 sampling points, and a sampling time interval of 0.1 μs.
[0039] The specific steps for identification are as follows: 1. A VLF-MF lightning electromagnetic wave waveform with a sampling rate of 10MHz and a sampling time window of 2ms was acquired using a lightning monitoring system. The timing waveform of this lightning wave is as follows: Figure 5 As shown; 2. Wavelet denoising is applied to the lightning electromagnetic wave, Hann windowing is used to suppress spectral leakage, and normalization is performed. The processed VLF-MF lightning electromagnetic wave waveform is as follows: Figure 6 As shown.
[0040] 3. Perform a fast Fourier transform on the preprocessed VLF-MF lightning electromagnetic wave to obtain a single-sided amplitude spectrum of 3kHz-3MHz.
[0041] 4. L2 normalization is performed on the amplitude spectrum vector to eliminate the influence of amplitude difference, and the VLF-MF lightning electromagnetic wave spectrum characteristics are obtained as shown in Figure 7 .
[0042] 5. The VLF-MF lightning waveform is compared with the pre-constructed four typical VLF-MF lightning waveform library one by one, and the FFT-cosine subsection similarity is calculated S b .
[0043] The 3kHz-3MHz frequency band is divided into three sub-bands: low frequency sub-band 3kHz-30kHz, medium frequency sub-band 30kHz-300kHz, and high frequency sub-band 300kHz-3MHz. The FFT-cosine subsection similarity of the three sub-bands is calculated by formula (1). The calculation results of the matching waveform are as follows: 1) Low frequency sub-band FFT-cosine subsection similarity S1=0.897 2) Medium frequency sub-band FFT-cosine subsection similarity S2=0.811 3) Medium frequency sub-band FFT-cosine subsection similarity S3=0.949 6. Calculate the FFT-cosine subsection fusion similarity S i .
[0044] The weighting value of the FFT-cosine subsection fusion similarity K b The weight selection is based on the propagation stability. Higher weight is set for the low frequency sub-band, and lower weight is given to the high frequency attenuation serious sub-band. Therefore, the low frequency sub-band weighting value K 1 is 0.7, the medium frequency sub-band weighting value K 2 is 0.2, and the high frequency sub-band weighting value K 3 is 0.1. The calculated FFT-cosine subsection similarity S b and the weighting value of the FFT-cosine subsection fusion similarity K b are substituted into formula (4) to calculate the FFT-cosine subsection fusion similarity S i , and the calculation results of the matching waveform are as follows: (10) Determine whether the FFT-cosine subsection fusion similarity S matches the preset threshold Sa, and quickly identify the VLF-MF lightning electromagnetic wave type. The identification preset threshold Sa According to the historical sample and the test set experimental result, 0.85 is selected, and the calculation result of the matching waveform S i greater than S a , and it is concluded that the recognition result is that the waveform is a preliminary breakdown (PB) electromagnetic wave.
[0045] The embodiment of the application also provides a spatial electromagnetic wave recognition device for an intelligent lightning positioning system, which comprises: A data acquisition and preprocessing module is configured to acquire VLF-MF lightning electromagnetic waves to be recognized, and perform noise reduction and window function processing on the waveform of the VLF-MF lightning electromagnetic waves; perform fast Fourier transform on the VLF-MF lightning electromagnetic waves to obtain the spectral characteristics of the VLF-MF lightning electromagnetic waves; and perform L2 normalization processing on the spectral characteristics of the VLF-MF lightning electromagnetic waves. A lightning waveform library construction module is configured to pre-construct a lightning waveform library containing four types of typical VLF-MF lightning electromagnetic waves, i.e., return stroke (RS), preliminary breakdown (PB), narrow bipolar event (NBE), and intracloud lightning (IC). A similarity calculation module is configured to compare the normalized spectral characteristics with the pre-constructed lightning waveform library one by one to determine FFT-cosine subsection similarity. A fusion module is configured to fuse all FFT-cosine subsection similarities to obtain FFT-cosine subsection fusion similarity. A matching module is configured to determine the lightning electromagnetic wave type of the VLF-MF lightning electromagnetic wave according to the matching result of the FFT-cosine subsection fusion similarity and a preset threshold.
[0046] The embodiment of the application also provides a computer device, which comprises a memory and a processor, the memory stores a computer program, and the computer program is executed by the processor to make the processor execute the following steps: Step 101: acquiring VLF-MF lightning electromagnetic waves to be recognized, and performing noise reduction and window function processing on the waveform of the VLF-MF lightning electromagnetic waves; Step 102: performing fast Fourier transform on the VLF-MF lightning electromagnetic waves to obtain the spectral characteristics of the VLF-MF lightning electromagnetic waves; Step 103, L2 normalization processing is performed on the spectrum characteristics of the VLF-MF lightning electromagnetic wave; Step 104, a lightning waveform library containing four typical VLF-MF lightning electromagnetic waves of a ground flash return stroke (RS), a preliminary breakdown (PB), a narrow bipolar event (NBE) and an intracloud lightning (IC) is constructed in advance; Step 105, FFT-cosine segment similarity is determined by comparing the normalized spectrum characteristics with the lightning waveform library constructed in advance one by one; Step 106, FFT-cosine segment fusion similarity is obtained by fusing all FFT-cosine segment similarities; Step 107, the lightning electromagnetic wave type of the VLF-MF lightning electromagnetic wave is determined according to the FFT-cosine segment fusion similarity and a preset threshold Matching results of .
[0047] The embodiment of the application also provides a computer readable storage medium storing a computer program, and the computer program is executed by a processor to make the processor execute the following steps: Step 101, a VLF-MF lightning electromagnetic wave to be identified is acquired, and waveform of the VLF-MF lightning electromagnetic wave is denoised and window function processed; Step 102, fast Fourier transform is performed on the VLF-MF lightning electromagnetic wave to obtain spectrum characteristics of the VLF-MF lightning electromagnetic wave; Step 103, L2 normalization processing is performed on the spectrum characteristics of the VLF-MF lightning electromagnetic wave; Step 104, a lightning waveform library containing four typical VLF-MF lightning electromagnetic waves of a ground flash return stroke (RS), a preliminary breakdown (PB), a narrow bipolar event (NBE) and an intracloud lightning (IC) is constructed in advance; Step 105, FFT-cosine segment similarity is determined by comparing the normalized spectrum characteristics with the lightning waveform library constructed in advance one by one; Step 106, FFT-cosine segment fusion similarity is obtained by fusing all FFT-cosine segment similarities; Step 107, determining whether the similarity of the FFT-cosine segment fusion is greater than a preset threshold value Matching results of determining a lightning electromagnetic wave type of the VLF-MF lightning electromagnetic wave.
[0048] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by a computer program instructing related hardware, and the program can be stored in a non-volatile computer readable storage medium. When the program is executed, it can include the processes of the above-mentioned embodiment methods. Any reference to memory, storage, database or other medium used in the embodiments provided by the present application 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. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.
[0049] The technical features of the above embodiments can be combined in any way. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described, but as long as the combinations of the technical features do not exist contradictory, they should be considered as the scope of the present application.
[0050] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are all within the scope of protection of the present application. Therefore, the scope of protection of the patent of the present application should be subject to the appended claims.
Claims
1. A spatial electromagnetic wave identification method for an intelligent lightning location system, characterized in that, The method includes: The VLF-MF lightning electromagnetic wave to be identified is acquired, and the waveform of the VLF-MF lightning electromagnetic wave is subjected to noise reduction and window function processing. The VLF-MF lightning electromagnetic wave is subjected to a fast Fourier transform to obtain the spectral characteristics of the VLF-MF lightning electromagnetic wave. The spectral characteristics of the VLF-MF lightning electromagnetic wave are subjected to L2 normalization. A lightning waveform library containing four typical VLF-MF lightning electromagnetic waves is pre-constructed, including Return Stroke (RS), Preliminary Breakdown (PB), Narrow Bipolar Event (NBE), and Intracloud Lightning (IC). The FFT-cosine segment similarity is determined by comparing the normalized spectral characteristics with a pre-built lightning waveform library one by one. The FFT-cosine segment similarity is obtained by fusing all FFT-cosine segment similarities. Based on the FFT-cosine segmentation fusion similarity and a preset threshold Matching results The lightning electromagnetic wave type of the VLF-MF lightning electromagnetic wave is determined.
2. The spatial electromagnetic wave identification method for an intelligent lightning location system according to claim 1, characterized in that, The L2 normalization processing of the spectral characteristics of the VLF-MF lightning electromagnetic wave specifically includes: according to , Determine the L2 norm of the spectral characteristics of the VLF-MF lightning electromagnetic wave respectively; L2 normalization is performed based on the L2 norm of the spectral characteristics of the VLF-MF lightning electromagnetic wave.
3. The spatial electromagnetic wave identification method for an intelligent lightning location system according to claim 1 or 2, characterized in that, The step of comparing the normalized spectral features with a pre-built lightning waveform library one by one to determine the FFT-cosine segment similarity specifically includes: according to Determine FFT-cosine segment similarity ;in, X The spectral vector of the VLF-MF lightning wave is... Y i The spectral vector of any VLF–MF lightning wave from a pre-built lightning waveform library.
4. The spatial electromagnetic wave identification method for an intelligent lightning location system according to claim 3, characterized in that, The process of fusing all FFT-cosine segment similarities to obtain FFT-cosine segment fusion similarity specifically includes: based on Obtain FFT-cosine fusion similarity ;in, The weighted value for the FFT-cosine segment fusion similarity. .
5. The spatial electromagnetic wave identification method for an intelligent lightning location system according to claim 4, characterized in that, The sampling rate of the VLF-MF lightning electromagnetic wave to be identified is consistent with the sampling rate of the VLF-MF lightning wave in the lightning waveform library.
6. The spatial electromagnetic wave identification method for an intelligent lightning location system according to claim 5, characterized in that, The step of determining the lightning electromagnetic wave type of the VLF-MF lightning electromagnetic wave based on the matching result of the FFT-cosine segmented fusion similarity and the preset threshold specifically includes: Compare the FFT-cosine segment fusion similarity. S i With preset threshold S a Size; like S i Less than S a If it does not match the current VLF-MF lightning wave in the lightning waveform library, continue to calculate the similarity with the next VLF-MF lightning wave in the lightning waveform library. like S i Greater than or equal to S a It matches the current VLF-MF lightning wave in the lightning waveform library and uses the lightning electromagnetic wave type of the current VLF-MF lightning wave as the identification result.
7. The spatial electromagnetic wave identification method for an intelligent lightning location system according to claim 6, characterized in that, The preset threshold S a Determined based on experimental results from historical samples or test sets.
8. A spatial electromagnetic wave identification device for an intelligent lightning location system, characterized in that, The device includes: The data acquisition and preprocessing module is used to acquire the VLF-MF lightning electromagnetic wave to be identified, and to perform noise reduction and window function processing on the waveform of the VLF-MF lightning electromagnetic wave; to perform fast Fourier transform on the VLF-MF lightning electromagnetic wave to obtain the spectral characteristics of the VLF-MF lightning electromagnetic wave; and to perform L2 normalization processing on the spectral characteristics of the VLF-MF lightning electromagnetic wave. The lightning waveform library construction module is used to pre-build a lightning waveform library containing four types of typical VLF-MF lightning electromagnetic waves: Return Stroke (RS), Preliminary Breakdown (PB), Narrow Bipolar Event (NBE), and Intracloud Lightning (IC). The similarity calculation module is used to compare the normalized spectral features with a pre-built lightning waveform library one by one to determine the FFT-cosine segment similarity. The fusion module is used to fuse all FFT-cosine segment similarities to obtain the FFT-cosine segment fusion similarity. The matching module is used to determine the lightning electromagnetic wave type of the VLF-MF lightning electromagnetic wave based on the matching result of the FFT-cosine segmented fusion similarity and the preset threshold.
9. A computer-readable storage medium storing a computer program that, when executed by a processor, causes the processor to perform the steps of the method as claimed in any one of claims 1 to 7.
10. A computer device comprising a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the steps of the method as claimed in any one of claims 1 to 7.