Dolphin whistle interpolation method based on linear prediction model, program, equipment and storage medium

Abstract

The method comprises the following steps: modeling a whistle signal based on a linear prediction model, estimating a parameter vector and a missing whistle vector based on an expectation maximization algorithm, calculating a secondary error of an estimation result, comparing the secondary error with a threshold value, and recovering the missing whistle signal through continuous iteration. The invention further provides a method for positioning the dolphin echo positioning signal in the original aliasing signal, a Gabor time frequency filter is used for filtering the aliasing signal, the position of a detection window where the click is located is determined, the problem that the correlation of a traditional matched filtering method is reduced is solved by adaptively updating a reference signal, the peak value of the dolphin echo positioning signal is enhanced through the TKEO, and the positioning accuracy of the dolphin echo positioning signal is improved. And the detection probability of the dolphin echo positioning signal in the aliasing signal is improved. According to the invention, the original whistle signal can be effectively recovered at the frequency and amplitude at the same time, and the higher harmonic component of the whistle can be recovered.

Description

Technical Field

[0001] The present invention belongs to the technical field of signal processing, and particularly relates to a dolphin whistle interpolation method, program, device and storage medium based on a linear prediction model. Background Art

[0002] Dolphin whistle signals are often used in underwater acoustic communication research, such as being used as carrier signals in bionic communication. However, the whistle signals recorded in the ocean are often contaminated by pulsed signals such as piling and clicks. Such pulsed signals have high energy and short duration, which will seriously affect the subsequent applications of the whistle signals. For example, in bionic communication, when extracting the spectral profile from the whistle signals obtained in the ocean to generate bionic communication signals, the subsequent processing steps of the whistle signals contaminated by pulsed signals are rather cumbersome, so they are often considered as unusable signals.

[0003] A simple method to eliminate the contamination of pulsed signals is to directly delete the interfering signals in the time domain, but this will cause missing whistle segments in the whistle signals and discontinuities in frequency and phase, affecting the subsequent applications of the whistle signals. Therefore, developing a signal interpolation algorithm to restore the missing whistle segments in the signals is an important task.

[0004] Patent CN115409062A discloses a method, system and storage medium for constructing bionic dolphin whistle signals. The obtained original dolphin whistle signals are subjected to short-time Fourier transform and adaptive profile extraction, and finally least-square polynomial fitting is performed to realize the modeling of bionic dolphin whistle signals. However, this method is applicable to the modeling of complete whistle signals and is not suitable for the restoration of shorter missing whistle segments. If this method is applied to the restoration of missing whistle segments, it is difficult to ensure the periodicity and phase continuity of the restored whistle segments. Patent CN118401935A discloses a sensor signal interpolation method. The non-equidistant interpolation method is used to calculate the measurement points for the measured signal, and after being converted into an electrical signal, it forms a data pair with the original measurement points, and polynomial interpolation is performed on the data pair. However, sensor signals are usually continuous and relatively stable, with certain regularity and predictability, while dolphin whistle signals are rapidly changing. Moreover, whistle signals are often interfered by pulsed signals such as piling and clicks in the marine environment. In the case of interference, the characteristics of the whistle signals may be masked. Therefore, simple polynomial interpolation may not be able to accurately restore the missing whistle signals. Summary of the Invention

[0005] The purpose of the present invention is to provide a dolphin whistle interpolation method, program, device and storage medium based on a linear prediction model, which can restore the whistle signals contaminated by pulsed signals such as piling and clicks.

[0006] A method for interpolating dolphin whistles based on a linear prediction model, comprising the following steps:

[0007] Step 1: Obtain the original aliased signal and locate the dolphin echolocation signal in the original aliased signal;

[0008] Step 2: Set the data at the positions of each dolphin echolocation signal in the original aliased signal to zero to eliminate pulse interference, obtaining a whistle signal with missing whistle segments;

[0009] Step 3: Model the whistle signal based on the linear prediction model. For each missing whistle segment, estimate the linear prediction coefficient vector and the missing whistle segment based on the expectation-maximization algorithm, and recover the missing whistle signal through iterative calculation.

[0010] Further, in step 3, modeling the whistle signal based on the linear prediction model is specifically as follows:

[0011] The whistle signal containing a missing whistle segment x l consists of N samples, {x0, x1,..., x N-1}}, and the missing whistle segment is a segment of whistle samples of length l starting from the k-th sample. The missing whistle segment x l is expressed as:

[0012] x l = {x k , x k+1 , …, x k+l-1} T ;

[0013] Modeling the whistle signal based on the linear prediction model is:

[0014]

[0015] Simplifying the above formula to:

[0016] e = x - Xa

[0017] where p is the order of the linear prediction model; a = {a1, a2,..., a p} T is the linear prediction coefficient vector; e = {e p , e p+1 ,..., e k , e k+1 ,..., e k+l-1 , e k+l ,..., e N-1} T is the error vector of the whistle signal.

[0018] Further, in step 3, based on the expectation maximization algorithm, the linear prediction coefficient vector a and the missing whistle segment x l are estimated, and the missing whistle segment x l is recovered by iteration. The specific method is as follows:

[0019] Step 3.1: Initialize the iteration number t = 1, and let x l (1) be a zero vector, then the elements in x(1) and X(1) are all known;

[0020] Step 3.2: Estimate the linear prediction coefficient vector a(t);

[0021] a(t) = [X T (t)X(t)] -1 X T (t)x(t)

[0022] Step 3.3: Construct the first coefficient matrix A1(t) and the second coefficient matrix A2(t) according to each linear prediction coefficient {a1(t), a2(t),..., a p (t)} in a(t);

[0023]

[0024] Among them, A1(t) is a (l + p)×l-dimensional matrix; A2(t) is a (l + p)×2p-dimensional matrix; Step 3.4: Estimate the missing whistle segment x l (t + 1);

[0025]

[0026] Among them, x r = {x k-p , x k-p+1 , …, x k-2 , x k-1 , x k+l , x k+l+1 , …, x k+l+p-1};

[0027] Step 3.5: Obtain x(t + 1) and X(t + 1) according to the missing whistle segment x l (t + 1);

[0028] Step 3.6: Calculate the estimation error e2(t);

[0029] e2(t) = x T (t + 1)x(t + 1) + a T (t)X T (t + 1)X(t + 1)a(t) - 2a T (t)X T(t + 1)x(t + 1)

[0030] Step 3.7: Compare e2(t) with a preset error threshold. If e2(t) is less than the error threshold, output the missing whistle segment x l (t + 1), and complete the missing recovery of the whistle signal; otherwise, let t = t + 1, and return to Step 3.2.

[0031] Furthermore, the method for locating the dolphin echolocation signal in the original aliased signal in Step 1 is specifically as follows:

[0032] Step 1.1: Perform a short-time Fourier transform on the original aliased signal to obtain the time-frequency diagram Y(t, f) corresponding to the original aliased signal;

[0033] Step 1.2: Perform two-dimensional Gabor filtering on the time-frequency diagram Y(t, f) in the vertical direction to obtain an enhanced image G(t, f) of the filtered dolphin echolocation signal;

[0034] Step 1.3: Calculate the mean value of G(t, f) corresponding to each moment t According to the threshold T0, take each continuous corresponding time range as a sliding detection window;

[0035] Step 1.4: Adopt an adaptive matching filtering method to intercept the dolphin echolocation signal in the first sliding detection window. For other sliding detection windows, use the dolphin echolocation signal intercepted in the previous sliding detection window as the reference signal for the current sliding detection window, and filter the original aliased signal in the current sliding detection window according to the reference signal to obtain the filtered signal y mi (n);

[0036] Step 1.5: Calculate TKEO for the filtered signal y mi (n), take the time corresponding to the peak value of TKEO in each sliding detection window as the pulse center, and locate the dolphin echolocation signal in the original aliased signal according to the pulse center.

[0037] Furthermore, the impulse response function of the two-dimensional Gabor filtering in Step 1.2 is:

[0038]

[0039] where λ is the wavelength of the filter; θ is the direction of the filter; σ is the standard deviation of the Gaussian function; t′ = t·cosθ + f·sinθ, f′ = -t·sinθ + f·cosθ;

[0040] Perform two-dimensional Gabor filtering on the time-frequency diagram Y(t,f) in the vertical direction to obtain the enhanced image G(t,f) of the filtered dolphin echolocation signal;

[0041] G(t,f) = Y(t,f) * g(t,f,λ,θ,σ t ,σ f )

[0042] where "*" represents the discrete convolution operation.

[0043] Further, in step 1.4, except for the first sliding detection window, the filtered signal y mi (n) of the i-th sliding detection window is:

[0044]

[0045] where y i (n) represents the original frequency-domain signal within the i-th sliding detection window; s * (n) is the complex conjugate of s(n); represents the dolphin echolocation signal detected according to step 1.5 within the (i - 1)-th sliding detection window; n0 is the time delay of s i (n).

[0046] Further, for the filtered signal y mi (n) in step 1.5, the specific calculation of TKEO is:

[0047]

[0048] A computer device / system, including a memory, a processor, and a computer program stored on the memory, where the processor executes the computer program to implement the steps of the above-mentioned dolphin whistle interpolation method based on a linear prediction model.

[0049] A computer-readable storage medium, on which a computer program / instructions are stored, and when the computer program / instructions are executed by a processor, the steps of the above-mentioned dolphin whistle interpolation method based on a linear prediction model are implemented.

[0050] A computer program product, including a computer program / instructions, and when the computer program / instructions are executed by a processor, the steps of the above-mentioned dolphin whistle interpolation method based on a linear prediction model are implemented.

[0051] The beneficial effects of the present invention are as follows:

[0052] The present invention models the whistle signal based on a linear prediction model, and estimates the parameter vector and the missing whistle vector based on the expectation maximization algorithm. Calculate the quadratic error for the estimation result and compare it with a threshold value, and continuously iterate to recover the missing whistle signal. Different from the traditional interpolation method that can only recover a small part of the whistle data in the spectral profile, the present invention can effectively recover the original whistle signal in both frequency and amplitude, and can also recover the high-order harmonic components of the whistle.

[0053] The present invention also provides a method for locating dolphin echolocation signals in the original aliased signal. In order to overcome the difficulty of click detection caused by high-energy whistles, apply a Gabor time-frequency filter to filter the aliased signal to determine the position of the detection window where the click is located; the present invention designs an AMF-TKEO algorithm, which solves the problem of the decrease in correlation of the traditional matching filtering method by adaptively updating the reference signal. At the same time, the application of the detection window also excludes the influence of local correlation peaks on click detection. The present invention not only provides a solution for click suppression in aliased signals, but can also be simply applied to click detection tasks under low signal-to-noise ratios. Compared with traditional methods, the present invention has higher accuracy and robustness, and greatly improves the detection probability of dolphin echolocation signals in aliased signals. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a schematic flow chart of recovering the missing whistle signal based on the linear prediction model in the present invention.

[0055] Figure 2 It is a spectrogram of the original dolphin whistle.

[0056] Figure 3 It is a spectrogram of the recovered dolphin whistle.

[0057] Figure 4 It is a detailed diagram of the recovered dolphin whistle.

[0058] Figure 5 It is a schematic flow chart of locating dolphin echolocation signals from the original aliased signal in the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0059] The present invention will be further described below with reference to the drawings.

[0060] Dolphin whistle signals are often used in underwater acoustic communication research. However, the whistle signals recorded in the ocean are often contaminated by pulsed signals such as piling and clicks. If the interference signals are directly deleted in the time domain, it will cause missing whistle segments in the whistle signal, and it is discontinuous in frequency and phase, affecting the subsequent application of the whistle signal. Therefore, developing a signal interpolation algorithm to recover the missing whistle segments in the signal is an important task.

[0061] Contains a missing whistle segment x l The whistle signal consists of N samples, {x0, x1,..., x N-1}, and the missing whistle segment is a segment of whistle samples of length l starting from the k-th sample. The missing whistle segment x l is expressed as:

[0062] x l ={x k , x k+1 ,…, x k+l-1} T ;

[0063] Model the whistle signal based on the linear prediction model as:

[0064]

[0065] Simplify the above formula to:

[0066] e = x - Xa

[0067] where p is the order of the linear prediction model; a = {a1, a2,..., a p} T is the linear prediction coefficient vector; e = {e p , e p+1 ,..., e k , e k+1 ,..., e k+l-1 , e k+l ,..., e N-1} T is the error vector of the whistle signal.

[0068] The parameter vector a and the missing whistle vector x l in the formula are unknown.

[0069] Based on the idea of the expectation maximization algorithm, in the first stage, by assuming that the missing whistle data x l = 0, obtain a sub-optimal estimate of the linear prediction coefficient vector a

[0070]

[0071] Construct the first coefficient matrix A1 and the second coefficient matrix A2 according to each linear prediction coefficient {a1, a2,..., a p} in a;

[0072]

[0073] Among them, A1 is a matrix of dimension (l + p)×l. Each column in A1 contains a sequence of consecutive {1, -a1, -a2,..., -a p} elements, and the first column starts with the element 1, and the last column ends with the element -a p ; A2(t) is a matrix of dimension (l + p)×2p. The elements in A2(t) can be understood to consist of four parts, namely O is a zero matrix,

[0074] Solve the error vector e of the whistle signal according to the first coefficient matrix A1 and the second coefficient matrix A2:

[0075] e = A1x l + A2x r

[0076] x l The least squares estimate of can be expressed as:

[0077]

[0078] Among them, is the least squares estimate value of the missing whistle data x l .

[0079] Calculate and The quadratic error of:

[0080] e T e = x T x + a T X T Xa - 2a T X T x

[0081] e T e is obviously a scalar. For the convenience of expression, let e2 = e T e, and compare it with the set error threshold. If it is less than the error threshold, the missing whistle signal can be solved by the least squares estimate . Otherwise, repeat the above steps until the quadratic error is less than the threshold to end, and the missing whistle signal can be obtained.

[0082] The above specific iterative calculation process includes the following steps:

[0083] Step 3.1: Initialize the iteration number t = 1, and let x l (1) be a zero vector, then the elements in x(1) and X(1) are all known;

[0084] Step 3.2: Estimate the linear prediction coefficient vector a(t);

[0085] a(t) = [X T (t)X(t)] -1 X T (t)x(t)

[0086] Step 3.3: Construct the first coefficient matrix A1(t) and the second coefficient matrix A2(t) according to the linear prediction coefficients {a1(t), a2(t),..., a p (t)};

[0087] Step 3.4: Estimate the missing whistle segment x l (t + 1);

[0088]

[0089] Step 3.5: Obtain x(t + 1) and X(t + 1) according to the missing whistle segment x l (t + 1);

[0090] Step 3.6: Calculate the estimation error e2(t);

[0091] e2(t) = x T (t + 1)x(t + 1) + a T (t)X T (t + 1)X(t + 1)a(t) - 2a T (t)X T (t + 1)x(t + 1)

[0092] Step 3.7: Compare e2(t) with the preset error threshold. If e2(t) is less than the error threshold, output the missing whistle segment x l (t + 1), and complete the missing recovery of the whistle signal; otherwise, let t = t + 1, and return to Step 3.2.

[0093] The present invention also provides a method for locating dolphin echolocation signals in the original aliased signal, as Figure 5 shown, including the following steps:

[0094] Step 1.1: Perform a short-time Fourier transform on the original aliased signal to obtain the time-frequency diagram Y(t, f) corresponding to the original aliased signal;

[0095] Step 1.2: Perform two-dimensional Gabor filtering on the time-frequency diagram Y(t, f) in the vertical direction to obtain the enhanced image G(t, f) of the filtered dolphin echolocation signal;

[0096] The impulse response function of the two-dimensional Gabor filtering is:

[0097]

[0098] Among them, λ is the wavelength of the filter; θ is the direction of the filter; σ is the standard deviation of the Gaussian function; t′ = t·cosθ + f·sinθ, f′ = -t·sinθ + f·cosθ;

[0099] Perform two-dimensional Gabor filtering on the time-frequency diagram Y(t, f) in the vertical direction to obtain the enhanced image G(t, f) of the filtered dolphin echolocation signal;

[0100] G(t, f) = Y(t, f) * g(t, f, λ, θ, σ t , σ f )

[0101] Among them, "*" represents the discrete convolution operation;

[0102] Step 1.3: Calculate the mean value of G(t, f) corresponding to each moment t According to the threshold T0, each continuous corresponding time range is used as a sliding detection window;

[0103] Step 1.4: Adopt the adaptive matched filtering method to intercept the dolphin echolocation signal in the first sliding detection window. For other sliding detection windows, use the dolphin echolocation signal intercepted in the previous sliding detection window as the reference signal for the current sliding detection window, and filter the original aliased signal in the current sliding detection window to obtain the filtered signal y mi (n);

[0104] Except for the first sliding detection window, the filtered signal y mi (n) for the i-th sliding detection window is:

[0105]

[0106] Among them, y i (n) represents the original frequency domain signal in the i-th sliding detection window; s * (n) is the complex conjugate of s(n); represents the dolphin echolocation signal detected according to Step 1.5 in the (i - 1)-th sliding detection window; n0 is the time delay of s i (n);

[0107] Step 1.5: Calculate the TKEO for the filtered signal y mi (n), take the time corresponding to the peak value of TKEO in each sliding detection window as the pulse center, and locate the dolphin echolocation signal in the original aliased signal according to the pulse center.

[0108]

[0109] TKEO is characterized by having instantaneous tracking ability with only three consecutive signal samples, so it can enhance the peak after matched filtering, thereby highlighting the position of the dolphin echolocation signal.

[0110] After enhancing the peak of the dolphin echolocation signal, the central position of the dolphin echolocation signal in each detection window is the position corresponding to the peak point in that window. Thus, the position coordinates of each dolphin echolocation signal in the original aliased signal can be represented.

[0111] Example 1:

[0112] Select a beluga whistle signal, which is aliased with a click signal containing 20 pulses. The whistle signal has a duration of 0.6 s, a sampling rate of 24000 Hz, a fundamental frequency of about 3500 Hz, a fundamental bandwidth of about 600 Hz, and has two obvious harmonics within the sampling frequency range. Its time-frequency spectrogram is as Figure 2 shown. Set the position of each click pulse to zero in the whistle signal aliased with the click pulse train to obtain a whistle signal with missing whistle segments.

[0113] The order p of the linear prediction model affects the recovery effect of the missing whistle signal. Through simulation comparison, it is found that the recovery effect of the missing whistle is the best when p = 20. Therefore, p is taken as 20 in this embodiment.

[0114] In this embodiment, the whistle signal consists of 14400 samples, including a missing whistle segment with a length l = 192. The time-frequency spectrogram of the dolphin whistle recovered by the interpolation method proposed by the present invention is as Figure 3 shown, and the comparison diagram of the time-domain details between the recovered whistle signal and the original whistle signal is as Figure 4 shown. It can be seen that the present invention more accurately recovers the time-domain waveform of the original whistle, and can recover the two harmonic components of the whistle, and more accurately completes the task of recovering the missing whistle while maintaining the time-frequency structure of the whistle component.

[0115] The above is only the preferred embodiment of the present invention and is not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A dolphin whistle interpolation method based on a linear prediction model, characterized in that, It includes the following steps: Step 1: Obtain the original aliased signal and locate the dolphin echolocation signal in the original aliased signal; Step 2: Set the data at the positions of each dolphin echolocation signal in the original aliased signal to zero to eliminate pulse interference, and obtain a whistle signal with missing whistle segments; Step 3: Model the whistle signal based on a linear prediction model. For each missing whistle segment, estimate the linear prediction coefficient vector and the missing whistle segment based on the expectation-maximization algorithm, and recover the missing whistle signal through iterative calculation.

2. The method for interpolating dolphin whistles based on a linear prediction model according to claim 1, characterized in that: In step 3, modeling the whistle signal based on the linear prediction model is specifically as follows: The whistle signal containing a missing whistle segment x l consists of N samples, {x0, x1,..., x N-1}, and the missing whistle segment is a segment of whistle samples of length l starting from the k-th sample. The missing whistle segment x l is expressed as: x l = {x k , x k+1 , …, x k+l-1} T ; Modeling the whistle signal based on the linear prediction model as: Simplify the above formula to: e = x - Xa where p is the order of the linear prediction model; a = {a1, a2,..., a p} T is the linear prediction coefficient vector; e = {e p , e p+1 ,..., e k , e k+1 ,..., e k+l-1 , e k+l ,..., e N-1} T is the error vector of the whistle signal.

3. The method for interpolating dolphin whistles based on a linear prediction model according to claim 2, wherein: In step 3, based on the expectation-maximization algorithm, estimate the linear prediction coefficient vector a and the missing whistle segment x l and recover the missing whistle segment x by iteration l The specific method is as follows: Step 3.1: Initialize the iteration number \(t = 1\), and let \(\mathbf{x}^{(1)}\) l be the zero vector, so both the elements of \(\mathbf{x}^{(1)}\) and \(\mathbf{X}^{(1)}\) are known; Step 3.2: Estimate the linear prediction coefficient vector a(t); a(t) = [X T (t)X(t)] -1 X T (t)x(t) Step 3.3: Construct a first coefficient matrix A1(t) and a second coefficient matrix A2(t) according to each linear prediction coefficient {a1(t), a2(t),..., a p (t)}; where, A1(t) is a matrix of dimension (l + p) × l; A2(t) is a matrix of dimension (l + p) × 2p; Step 3.4: Estimate the missing whistle segment x l (t + 1); where x r = {x k-p , x k-p+1 , …, x k-2 , x k-1 , x k+l , x k+l+1 , …, x k+l+p-1}; Step 3.5: According to the missing whistle segment x l (t + 1), obtain x(t + 1) and X(t + 1); Step 3.6: Calculate the estimated error e2(t); e2(t) = x T (t + 1)x(t + 1) + a T (t)X T (t + 1)X(t + 1)a(t) - 2a T (t)X T (t + 1)x(t + 1) Step 3.7: Compare e2(t) with a preset error threshold. If e2(t) is less than the error threshold, output the missing whistle segment x l (t + 1) to complete the missing recovery of the whistle signal; otherwise, set t = t + 1 and return to Step 3.

2.

4. A method for interpolating dolphin whistles based on a linear prediction model according to claim 1, characterized in that: The method for locating the dolphin echolocation signal in the original aliased signal in step 1 is specifically as follows: Step 1.1: Perform a short-time Fourier transform on the original aliased signal to obtain the time-frequency diagram Y(t, f) corresponding to the original aliased signal; Step 1.2: Perform two-dimensional Gabor filtering on the time-frequency diagram Y(t, f) in the vertical direction to obtain an enhanced image G(t, f) of the filtered dolphin echolocation signal; Step 1.3: Calculate the mean value of G(t,f) at each moment t According to the threshold T0, for each continuous corresponding time range as a sliding detection window; Step 1.4: Using the adaptive matched filtering method, intercept the dolphin echolocation signal within the first sliding detection window. For other sliding detection windows, use the dolphin echolocation signal intercepted within the previous sliding detection window as the reference signal for the current sliding detection window, and filter the original aliased signal within the current sliding detection window according to the reference signal to obtain the filtered signal y mi (n); Step 1.5: For the filtered signal y mi (n), calculate the TKEO, take the time corresponding to the peak value of TKEO in each sliding detection window as the pulse center, and locate the dolphin echolocation signal in the original aliased signal according to the pulse center.

5. A method for interpolating dolphin whistles based on a linear prediction model according to claim 4, characterized in that: The impulse response function of the two-dimensional Gabor filtering in step 1.2 is: where, λ is the wavelength of the filter; θ is the direction of the filter; σ is the standard deviation of the Gaussian function; t′ = t·cosθ + f·sinθ, f′ = -t·sinθ + f·cosθ; Perform two-dimensional Gabor filtering on the time-frequency diagram Y(t, f) in the vertical direction to obtain an enhanced image G(t, f) of the filtered dolphin echolocation signal; G(t,f) = Y(t,f) * g(t,f,λ,θ,σ t ,σ f ) where, "*” represents the discrete convolution operation.

6. The interpolation method for dolphin whistles based on a linear prediction model according to claim 4, characterized in that: In step 1.4, for the i-th sliding detection window except the first one, the filtered signal y mi (n) is as follows: where y i (n) represents the original frequency-domain signal within the i-th sliding detection window; s * (n) is the complex conjugate of s(n); represents the dolphin echolocation signal detected according to Step 1.5 within the (i - 1)-th sliding detection window; n0 is the time delay of s i (n).

7. A method for interpolating dolphin whistles based on a linear prediction model according to claim 6, characterized in that: For the filtered signal y in step 1.5 mi (n), the calculation of the TKEO is specifically as follows:

8. A computer device / equipment / system, comprising a memory, a processor, and a computer program stored on the memory, characterized in that: The processor executes the computer program to implement the steps of the method described in any one of claims 1 to 7.

9. A computer-readable storage medium having computer programs / instructions stored thereon, characterized in that: When the computer program / instructions are executed by the processor, the steps of the method described in any one of claims 1 to 7 are implemented.

10. A computer program product, comprising a computer program / instructions, characterized in that: When the computer program / instructions are executed by the processor, the steps of the method described in any one of claims 1 to 7 are implemented.

Images