Method for distinguishing electric shock current signal and non-electric shock current signal

Abstract

The invention discloses a method for distinguishing an electric shock current signal and a non-electric shock current signal, and the method comprises the following steps: carrying out the time-frequency transformation of a collected current signal, and obtaining a time-frequency complex matrix; extracting a time-frequency characteristic quantity based on the time-frequency complex matrix; a threshold value is set according to the time-frequency characteristic quantity, and an electric shock current signal and a non-electric shock current signal are distinguished; calculating a signal energy characteristic quantity based on the time-frequency complex matrix; and a threshold value is set according to the signal energy characteristic quantity, and an electric shock current signal and a non-electric shock current signal are distinguished. According to the method, time-frequency conversion is carried out on the current signal, the time-frequency characteristic quantity and the signal energy characteristic quantity are extracted, comprehensive judgment is carried out according to the set threshold value, multiple kinds of characteristic information of the electric shock current signal in time and frequency are fully considered, the electric shock current signal can be recognized more accurately, the probability of misjudgment and missed judgment is effectively reduced, and the detection accuracy is improved. And the accuracy and reliability of electric shock detection are improved.

Description

Technical Field

[0001] The present invention relates to the field of electric shock current, and in particular to a method for distinguishing electric shock current signals from non-electric shock current signals. Background Art

[0002] In the wide application of modern power systems and electrical equipment, electrical safety issues have always been the focus of much attention. Electric shock accidents, as a common and dangerous electrical safety accident, may pose a serious threat to the safety of personnel's lives, and may also cause secondary disasters such as damage to electrical equipment and fire. Therefore, how to accurately and timely detect and distinguish electric shock current signals from non-electric shock current signals is of vital importance to ensure personal safety and the stable operation of electrical systems.

[0003] Electric shock current signals usually have some unique characteristics. For example, at the moment of electric shock, the current may change suddenly, including a sharp increase or decrease in amplitude, a sudden change in phase, etc. In addition, the distribution of electric shock current signals in time and frequency may also be significantly different from normal non-electric shock current signals. However, in actual electrical environments, current signals are often interfered by various factors, such as electromagnetic interference, load changes, power grid fluctuations, etc. These interference factors make it very difficult to accurately detect and distinguish electric shock current signals.

[0004] The traditional electric shock current signal detection method is mainly based on the monitoring and judgment of single characteristics such as current amplitude and phase. For example, by setting a fixed current threshold, when the current amplitude exceeds the threshold, it is judged that an electric shock accident has occurred. However, this method has obvious limitations. On the one hand, the fixed threshold cannot adapt to the current changes in different electrical equipment and environments, and it is easy to have misjudgment or missed judgment. For example, when some high-power electrical equipment is started, the current amplitude may exceed the set threshold instantaneously, but this does not necessarily mean that an electric shock accident has occurred; on the other hand, this method only considers the amplitude characteristics of the current, ignoring other important information of the current signal in time and frequency. For some complex electric shock situations, such as indirect contact and leakage, it may not be accurately detected.

[0005] In order to overcome the limitations of traditional detection methods, time-frequency analysis technology has gradually been applied to the detection and differentiation of electric shock current signals in recent years. Time-frequency analysis technology can analyze signals in both time and frequency dimensions at the same time, revealing the frequency components and energy distribution of signals at different times, so as to obtain the characteristic information of signals more comprehensively. For example, short-time Fourier transform (STFT) is a commonly used time-frequency analysis method, which performs Fourier transform on the signal at different time segments by windowing the signal to obtain the time-frequency distribution of the signal. However, the window size and shape of STFT are fixed and cannot be adjusted adaptively, so the analysis effect is not ideal for some non-stationary signals. Summary of the invention

[0006] The object of the present invention is to provide a method for distinguishing electric shock current signals from non-electric shock current signals, so as to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A method for distinguishing an electric shock current signal from a non-electric shock current signal comprises the following steps:

[0009] Step S1, performing time-frequency transformation on the collected current signal to obtain a time-frequency complex matrix;

[0010] Step S2, extracting time-frequency features based on the time-frequency complex matrix;

[0011] Step S3, setting a threshold value according to the time-frequency feature quantity to distinguish between electric shock current signals and non-electric shock current signals;

[0012] Step S4, calculating the signal energy characteristic quantity based on the time-frequency complex matrix;

[0013] Step S5, setting a threshold value according to the signal energy characteristic quantity to distinguish between electric shock current signals and non-electric shock current signals.

[0014] In the present invention, the time-frequency transform is an S transform, and the S transform of the signal h(t) is defined as:

[0015]

[0016] Where t is time, f is frequency, j is the imaginary unit, and τ is the center position of the Gaussian window function;

[0017] The inverse transform of the S transform is:

[0018]

[0019] The discrete form of the S transform is:

[0020]

[0021] Where H[K] is the discrete Fourier transform of the time series h(m), that is,

[0022]

[0023] The result of the S-transform of the signal is a two-dimensional time-frequency complex matrix, where the rows correspond to frequency and the columns correspond to time;

[0024] n=0 corresponds to the DC component of the signal, which is directly obtained from the mean value of the signal. n≠0 corresponds to other frequency components of the signal, which are processed as follows:

[0025] Step 101, calculating the fast Fourier transform H[K] of the time series h(m);

[0026] Step 102, let n=1;

[0027] Step 103, calculate the fast Fourier transform G[n,k] of the Gaussian window at the frequency sampling point n,

[0028] Step 104, shift the spectrum H[n] to H[k+n];

[0029] Step 105, calculate B[n,k], B[n,k]=H[K+n]*G[n,k];

[0030] Step 106, calculate the inverse fast Fourier transform of B[n,k] to obtain the S transform spectrum of the sampling frequency point n and the actual frequency n / NT, corresponding to the n+1th row of the S transform matrix;

[0031] Step 107, let n=n+1, repeat the above steps of calculating the fast Fourier transform of the Gaussian window at the frequency sampling point n to calculating the inverse fast Fourier transform, until the Nyquist frequency n=N / 2, the calculation is completed.

[0032] In the present invention, extracting time-frequency features based on the time-frequency complex matrix specifically includes:

[0033] Taking the module of the time-frequency complex matrix, we get the module time-frequency matrix, which is written as:

[0034]

[0035] Wherein, time sampling point m corresponds to sampling time mT, and frequency sampling point n corresponds to frequency n / NT;

[0036] Define the eigenvector:

[0037] A k [m]=A[k,m],m=0~N-1

[0038] Use the mean and standard deviation statistics to characterize the feature vector A k The characteristic information of [m] is expressed as follows:

[0039]

[0040] In the present invention, a threshold is set according to the time-frequency feature quantity to distinguish between electric shock current signals and non-electric shock current signals, specifically including:

[0041] Calculate the fundamental frequency-amplitude vector amplitude difference F1. When F1≥0.005, it is determined to be an electric shock current signal.

[0042] Calculate the fundamental frequency-amplitude vector standard deviation F2. When F2 ≥ 0.003, it is determined to be an electric shock current signal.

[0043] The custom standard deviation F3 of the fundamental frequency-amplitude vector is calculated. When F3 ≥ 0.008, it is determined to be an electric shock current signal.

[0044] In the present invention, the signal energy feature quantity is calculated based on the time-frequency complex matrix, specifically including:

[0045] According to Parsel's theorem, after time-frequency transformation, the energy of the signal within the sampling interval is expressed as:

[0046]

[0047] Ignoring high-frequency interference, the information in the 0-140Hz frequency band is selected as the effective information of the electric shock current signal, and the energy of the low-frequency band of the signal is expressed as:

[0048]

[0049] In the present invention, a threshold is set according to the signal energy characteristic quantity to distinguish between electric shock current signals and non-electric shock current signals, specifically including:

[0050] Calculate the low-frequency energy of the signal at any sampling time. The sampling time is 0 to N-1. The low-frequency energy of the signal forms a vector E L , define E minL =min(E L ) is the minimum value of the vector L, and the custom standard deviation F4 of the low-frequency energy vector is calculated;

[0051] When F4≥0.013, it is judged as an electric shock current signal.

[0052] In the present invention, the Gaussian window function is symmetric in time and frequency, and its Fourier transform is still a Gaussian function, and the Gaussian function has no side lobes, so that the local maximum value of the absolute value of the S transform is real.

[0053] In the present invention, the width and height of the Gaussian window in the S transform vary with the frequency, thus overcoming the defect that the height and width of the short-time Fourier transform window are fixed.

[0054] In the present invention, when the frequency f in the eigenvector corresponds to the fundamental frequency of the signal, A f [m] corresponds to the fundamental frequency-amplitude vector.

[0055] In the present invention, the custom standard deviation F3 is calculated using the minimum value as a benchmark instead of the mean value, so as to amplify the gap between different signals.

[0056] Compared with the prior art, the present invention has the following beneficial effects:

[0057] 1. The present invention performs time-frequency transformation on the current signal, extracts the time-frequency characteristic quantity and the signal energy characteristic quantity, and makes a comprehensive judgment according to the set threshold value, fully considering the various characteristic information of the electric shock current signal in time and frequency, and can more accurately identify the electric shock current signal, effectively reduce the probability of misjudgment and missed judgment, and improve the accuracy and reliability of electric shock detection; the S transform is used as the time-frequency transformation method, and the width and height of the Gaussian window change with the frequency, which overcomes the defect of the short-time Fourier transform window being fixed. The adaptive characteristics enable the method to better adapt to the complex changes of the electric shock current signal under different electrical environments, more accurately capture the characteristics of the signal, and further improve the accuracy of detection;

[0058] 2. The present invention not only considers the amplitude difference, standard deviation, custom standard deviation and other characteristics of the fundamental frequency-amplitude vector through in-depth analysis of the time-frequency complex matrix, but also analyzes the energy change characteristics of the signal, which can comprehensively reflect the characteristics of the electric shock current signal, thereby enhancing the detection capability of various complex electric shock situations; the actual current signal is easily interfered by various factors such as electromagnetic interference, load changes, and power grid fluctuations. The present invention extracts the low-frequency energy of the signal as effective information and calculates the custom standard deviation of the low-frequency energy vector, which can eliminate the influence of high-frequency interference and other noises to a certain extent, more accurately identify the electric shock current signal, and ensure that the electric shock situation can be reliably detected in a complex electrical environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is a schematic diagram of the S transform time-frequency of the present invention;

[0060] Figure 2 It is a schematic diagram for comparing the S-transformation results of the present invention;

[0061] Figure 3 This is a schematic diagram of the statistical results of the feature quantity F1 of the present invention;

[0062] Figure 4This is a schematic diagram of the statistical results of the characteristic quantity F2 of the present invention;

[0063] Figure 5 This is a schematic diagram of the statistical results of the characteristic quantity F3 of the present invention;

[0064] Figure 6 The energy curve of the current signal in the low frequency band changes with time in the present invention;

[0065] Figure 7 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0066] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0067] The present invention provides a technical solution:

[0068] A method for distinguishing an electric shock current signal from a non-electric shock current signal comprises the following steps:

[0069] Step S1, performing time-frequency transformation on the collected current signal to obtain a time-frequency complex matrix;

[0070] The time-frequency transform is an S transform, and the S transform of the signal h(t) is defined as:

[0071]

[0072] Where t is time, f is frequency, j is the imaginary unit, and τ is the center position of the Gaussian window function;

[0073] The Gaussian window function is symmetrical in time and frequency, and its Fourier transform is still a Gaussian function, and the Gaussian function has no side lobes, so that the local maximum of the absolute value of the S transform is real. The width and height of the Gaussian window in the S transform vary with the frequency, overcoming the defect that the height and width of the short-time Fourier transform window are fixed.

[0074] The inverse transform of the S transform is:

[0075]

[0076] The discrete form of the S transform is:

[0077]

[0078] Where H[K] is the discrete Fourier transform of the time series h(m), that is,

[0079]

[0080] The result of the S-transform of the signal is a two-dimensional time-frequency complex matrix, where the rows correspond to frequency and the columns correspond to time;

[0081] n=0 corresponds to the DC component of the signal, which is directly obtained from the mean value of the signal. n≠0 corresponds to other frequency components of the signal, which are processed as follows:

[0082] Step 101, calculating the fast Fourier transform H[K] of the time series h(m);

[0083] Step 102, let n=1;

[0084] Step 103, calculate the fast Fourier transform G[n,k] of the Gaussian window at the frequency sampling point n,

[0085] Step 104, shift the spectrum H[n] to H[k+n];

[0086] Step 105, calculate B[n,k], B[n,k]=H[K+n]*G[n,k];

[0087] Step 106, calculate the inverse fast Fourier transform of B[n,k] to obtain the S transform spectrum of the sampling frequency point n and the actual frequency n / NT, corresponding to the n+1th row of the S transform matrix;

[0088] Step 107, let n=n+1, repeat the above steps of calculating the fast Fourier transform of the Gaussian window at the frequency sampling point n to calculating the inverse fast Fourier transform, until the Nyquist frequency n=N / 2, the calculation is completed;

[0089] The electric shock current signal collected from the scene is a two-dimensional plane signal of amplitude-time. After S-transformation, a three-dimensional spatial surface of amplitude-time-frequency is obtained, which greatly enriches the signal characteristics.

[0090] Step S2, extracting time-frequency features based on the time-frequency complex matrix, specifically includes:

[0091] Taking the module of the time-frequency complex matrix, we get the module time-frequency matrix, which is written as:

[0092]

[0093] In the formula, the time sampling point m corresponds to the sampling time mT, and the frequency sampling point n corresponds to the frequency n / NT; Figure 1 As shown, the current signal contains the 3rd and 5th harmonics The 3D representation of the modulus time-frequency A[n,m] after the S transform, the amplitude-frequency envelope, is obtained by making a section on A[n,m] at any sampling time, showing the relationship between the amplitude and frequency at this moment; the amplitude-time envelope, is obtained by making a section on any frequency sampling point in A[n,m], showing the relationship between the amplitude of the corresponding frequency and time.

[0094] Therefore, the eigenvector A can be defined k [m]:

[0095] A k [m]=A[k,m],m=0~N-1

[0096] That is, the amplitude-time envelope with frequency k / NT is extracted from A[n,m] to represent the relationship between amplitude and time. In particular, when the frequency f corresponds to the fundamental frequency of the signal, A[n,m] corresponds to the fundamental frequency-amplitude vector. The feature vector A is characterized by two statistics: mean and standard deviation. k The characteristic information of [m] is expressed as follows:

[0097]

[0098] Step S3, setting a threshold value according to the time-frequency feature quantity to distinguish between electric shock current signals and non-electric shock current signals;

[0099] like Figure 2 As shown, after the electric shock current signal and the non-electric shock current signal are transformed by S, it can be seen that the amplitude of the non-electric shock current signal changes very little over time, the amplitude of the pig electric shock current signal changes greatly over time, and there is an obvious process of amplitude increase, which is consistent with the "characteristic of increasing amplitude" of the electric shock current signal in the time domain.

[0100] There are 68 groups of current signals collected in the current experiment, including 40 groups of live electric shock current signals: the first 15 groups are chickens and ducks being electrocuted, and the last 25 groups are pigs being electrocuted; 28 groups of non-electric shock current signals, 3 groups of resistance grounding fault currents; 14 groups of normal working currents; and 11 groups of arc fault currents. In order to distinguish between electric shock current signals and non-electric shock current signals, it is necessary to set the threshold. In the process of determining the threshold, the following statistical feature quantities are used.

[0101] The fundamental frequency-amplitude vector is A f [m], define the maximum value of the fundamental frequency-amplitude eigenvector as A maxf , the minimum value is A minf , the mean is The standard deviation is S f .

[0102] Specifically include:

[0103] like Figure 3 As shown, according to F1 = A maxf -A minf Calculate the fundamental frequency-amplitude vector amplitude difference F1. When F1≥0.005, it is determined to be an electric shock current signal. The electric shock current signal has the characteristic of increasing amplitude in the initial 2 to 3 cycles, which determines that the amplitude of the electric shock current signal must change with the passage of time, and there is a large difference between the maximum and minimum amplitudes. However, the non-electric shock current signal does not have this characteristic. Theoretically, it should be a straight line, and the amplitude difference must be small.

[0104] like Figure 4 As shown, according to Calculate the standard deviation F2 of the fundamental frequency-amplitude vector. When F2≥0.003, it is determined to be an electric shock current signal. The standard deviation is the best measurement indicator to measure whether a signal has fluctuations.

[0105] like Figure 5 As shown, according to Calculate the custom standard deviation F3 of the fundamental frequency-amplitude vector. When F3 ≥ 0.008, it is determined to be an electric shock current signal. Use the minimum value as the benchmark to replace the mean value for custom standard deviation calculation. Compared with F2, the value of F3 must increase, thereby amplifying the gap between different signals.

[0106] Table 1 Summary of time-frequency features

[0107]

[0108] The above Table 1 lists in detail the characteristic value ranges of electric shock current signals and non-electric shock current signals. It can be seen that when F1=0.005, or F2=0.003, or F3=0.008, the electric shock current signal and non-electric shock current signal can be completely distinguished.

[0109] Step S4, calculating the signal energy characteristic based on the time-frequency complex matrix; the electric shock current signal is a special signal with an increasing amplitude over time, and this amplitude change must be reflected in the S transformation matrix and the module matrix after S transformation. In other words, the module matrix information can fully show the particularity of the electric shock current signal.

[0110] Specifically include:

[0111] According to Parsel's theorem, the total energy of a signal is equal to the sum of the energies of each frequency. After S-transformation, assuming there is no spectrum leakage, the energy of the signal within a certain sampling interval can be expressed as:

[0112]

[0113] In other words, within a certain sampling interval, the energy of the signal is equal to the sum of the squares of the corresponding column elements of the modulus matrix.

[0114] From the Fourier analysis and wavelet packet analysis of the electric shock current signal, it can be seen that the high-frequency part of the electric shock current signal only has high-frequency interference but no effective information. Therefore, when obtaining the energy of the signal at a certain sampling time, the high-frequency interference can also be ignored. The length of the electric shock current signal to be analyzed is 5 cycles. From the basic principle and implementation of the S transform, it can be seen that the frequency resolution of the signal is 1 / NT, where T = 1 / Fs, Fs is the sampling rate, the number of sampling points per cycle N1 = Fs / Fb = Fs / 50, and the total number of sampling points N = 5N1, that is, the frequency resolution of the signal is 10Hz. The information in the 0-140Hz frequency band is selected as the effective information of the electric shock current signal, and the energy of the low-frequency band of the signal is expressed as:

[0115]

[0116] Step S5, setting a threshold value according to the signal energy characteristic quantity to distinguish between electric shock current signals and non-electric shock current signals;

[0117] Specifically include:

[0118] like Figure 6 As shown, the low-frequency energy of the signal at any sampling time can be obtained by It is found that the low-frequency energy of the signal in the sampling time segment 0 to N-1 constitutes the vector E L , combined with the diagram, it is obvious that the low-frequency energy of the non-electric shock current signal changes very little over time, while the electric shock current signal is not the case. Definition E minL =min(E L ), is the minimum value of vector L. Then the custom standard deviation of low-frequency energy vector F4:

[0119]

[0120] Table 2 Analysis results of energy vector custom standard deviation F4

[0121]

[0122] The characteristic quantity F4 analysis was performed on the existing 68 sets of data, including 40 sets of electric shock current signals and 28 sets of non-electric shock current signals, and the results are shown in Table 2. Obviously, when F4=0.013, the electric shock current signal and the non-electric shock current signal can be completely distinguished.

[0123] The S transform is used to analyze the current signal. The basic principle and implementation method of the S transform are described in detail. The electric shock current signal and the non-electric shock current signal are analyzed from two different perspectives: the time-frequency characteristics of the S transform and the energy change. On this basis, four feature quantities F1 to F4 are proposed. After actual data verification, each feature quantity can find a suitable threshold, such as F1 = 0.005, F2 = 0.003, F3 = 0.008, F4 = 0.013. A single feature quantity can completely distinguish the electric shock current signal from the non-electric shock current signal. The S transform and the Fourier transform and wavelet packet transform used in the previous article have great advantages. When Fourier transform and wavelet packet transform are used to analyze the electric shock current signal and the non-electric shock current signal, the analysis results are not ideal because the amplitude increase characteristics of individual electric shock current signals are not obvious. Only qualitative analysis can be performed, and it is concluded that the characteristic information of the electric shock current signal is hidden in the low-frequency band of about 50Hz. The S transform can quantitatively distinguish the electric shock current signal from the non-electric shock current signal, which fully demonstrates the advantages of the S transform in analyzing such signals.

[0124] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0125] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for distinguishing electric shock current signals from non-electric shock current signals, characterized in that: The following steps are involved: Step S1, performing time-frequency transformation on the collected current signal to obtain a time-frequency complex matrix; Step S2, extracting time-frequency features based on the time-frequency complex matrix; Step S3, setting a threshold value according to the time-frequency feature quantity to distinguish between electric shock current signals and non-electric shock current signals; Step S4, calculating the signal energy characteristic quantity based on the time-frequency complex matrix; Step S5, setting a threshold value according to the signal energy characteristic quantity to distinguish between electric shock current signals and non-electric shock current signals.

2. A method for distinguishing electric shock current signals from non-electric shock current signals according to claim 1, characterized in that: The time-frequency transform is an S transform, and the S transform of the signal h(t) is defined as: Where t is time, f is frequency, j is the imaginary unit, and τ is the center position of the Gaussian window function; The inverse transform of the S transform is: The discrete form of the S transform is: Where H[K] is the discrete Fourier transform of the time series h(m), that is, The result of the S-transform of the signal is a two-dimensional time-frequency complex matrix, where the rows correspond to frequency and the columns correspond to time; n=0 corresponds to the DC component of the signal, which is directly obtained from the mean value of the signal. n≠0 corresponds to other frequency components of the signal, which are processed as follows: Step 101, calculating the fast Fourier transform H[K] of the time series h(m); Step 102, let n=1; Step 103, calculate the fast Fourier transform G[n,k] of the Gaussian window at the frequency sampling point n, Step 104, shift the spectrum H[n] to H[k+n]; Step 105, calculate B[n,k], B[n,k]=H[K+n]*G[n,k]; Step 106, calculate the inverse fast Fourier transform of B[n,k] to obtain the S transform spectrum of the sampling frequency point n and the actual frequency n / NT, corresponding to the n+1th row of the S transform matrix; Step 107, let n=n+1, repeat the above steps of calculating the fast Fourier transform of the Gaussian window at the frequency sampling point n to calculating the inverse fast Fourier transform, until the Nyquist frequency n=N / 2, the calculation is completed.

3. A method for distinguishing electric shock current signals from non-electric shock current signals according to claim 1, characterized in that: Extracting the time-frequency feature quantity based on the time-frequency complex matrix specifically includes: Taking the module of the time-frequency complex matrix, we get the module time-frequency matrix, which is written as: Wherein, time sampling point m corresponds to sampling time mT, and frequency sampling point n corresponds to frequency n / NT; Define the eigenvector: A k [m]=A[k,m],m=0~N-1 Use the mean and standard deviation statistics to characterize the feature vector A k The characteristic information of [m] is expressed as follows:

4. The method for distinguishing electric shock current signals from non-electric shock current signals according to claim 1, characterized in that: Setting a threshold according to the time-frequency feature quantity to distinguish electric shock current signals from non-electric shock current signals specifically includes: Calculate the fundamental frequency-amplitude vector amplitude difference F1. When F1≥0.005, it is determined to be an electric shock current signal. Calculate the fundamental frequency-amplitude vector standard deviation F2. When F2 ≥ 0.003, it is determined to be an electric shock current signal. The custom standard deviation F3 of the fundamental frequency-amplitude vector is calculated. When F3 ≥ 0.008, it is determined to be an electric shock current signal.

5. The method for distinguishing electric shock current signals from non-electric shock current signals according to claim 1, characterized in that: Calculating the signal energy feature quantity based on the time-frequency complex matrix specifically includes: According to Parsel's theorem, after time-frequency transformation, the energy of the signal within the sampling interval is expressed as: Ignoring high-frequency interference, the information in the 0-140Hz frequency band is selected as the effective information of the electric shock current signal, and the energy of the low-frequency band of the signal is expressed as:

6. A method for distinguishing electric shock current signals from non-electric shock current signals according to claim 1, characterized in that: Setting a threshold value according to the signal energy characteristic quantity to distinguish electric shock current signals from non-electric shock current signals specifically includes: Calculate the low-frequency energy of the signal at any sampling time. The sampling time is 0 to N-1. The low-frequency energy of the signal forms a vector E L , define E minL =min(E L ) is the minimum value of the vector L, and the custom standard deviation F4 of the low-frequency energy vector is calculated; When F4≥0.013, it is judged as an electric shock current signal.

7. A method for distinguishing electric shock current signals from non-electric shock current signals according to claim 2, characterized in that: The Gaussian window function is symmetric in time and frequency, and its Fourier transform is still a Gaussian function, and the Gaussian function has no side lobes, so that the local maximum of the absolute value of the S transform is real.

8. A method for distinguishing electric shock current signals from non-electric shock current signals according to claim 2, characterized in that: The width and height of the Gaussian window in the S transform vary with the frequency, thus overcoming the defect that the height and width of the short-time Fourier transform window are fixed.

9. A method for distinguishing electric shock current signals from non-electric shock current signals according to claim 3, characterized in that: When the frequency f in the eigenvector corresponds to the fundamental frequency of the signal, A f [m] corresponds to the fundamental frequency-amplitude vector.

10. The method for distinguishing electric shock current signals from non-electric shock current signals according to claim 4, characterized in that: The custom standard deviation F3 is calculated using the minimum value as a benchmark instead of the mean value, and is used to amplify the gap between different signals.

Images