TEOAE signal detection method and system based on adaptive multi-scale wavelet transform

Abstract

The invention discloses a TEOAE signal detection method and system based on adaptive multi-scale wavelet transform, relates to the technical field of TEOAE signal detection, and solves the problems of poor signal denoising and feature extraction effects and limited signal detection accuracy and efficiency caused by adoption of a fixed wavelet basis function in the prior art. The method comprises the steps that TEOAE signals are preprocessed, key features are extracted, an optimal wavelet basis function is selected according to the key features, wavelet transformation is carried out, boundary effect processing is carried out, meanwhile, the detection effect is monitored in real time, parameters are adjusted, the accuracy and adaptability of signal detection are improved, the transient features of the signals can be better captured, and the detection accuracy of the TEOAE signals is improved. And the detection process can be optimized, and the adaptability and robustness of the TEOAE signal detection method can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of TEOAE signal detection, and in particular to a TEOAE signal detection method and system based on adaptive multi-scale wavelet transform. Background Art

[0002] TEOAE, or transient evoked otoacoustic emissions, is a type of otoacoustic emission (OAE). TEOAE refers to the audio energy released in a certain form after a certain latency period when the outer hair cells of the cochlea are stimulated by short-term sound (such as a click). This energy reflects the activity of active biophysical mechanisms in the cochlea and can be recorded in the external auditory canal.

[0003] TEOAE signal detection technology is based on the audio energy generated by the outer hair cells of the cochlea after transient sound stimulation. This energy is conducted through the middle ear structure, released into the external auditory canal, and can be recorded by specialized equipment. By analyzing these signals, the functional status of the cochlea can be understood.

[0004] Existing TEOAE signal detection technology usually uses fixed wavelet basis functions, which lacks adaptability to signal characteristics. The existing technical background mainly involves the application of wavelet transform in signal processing and some basic wavelet basis selection methods, which have the following disadvantages:

[0005] 1. Fixed wavelet basis may lead to poor effects of signal denoising and feature extraction;

[0006] 2. It is unable to dynamically adapt to the instantaneous changes of the signal, resulting in limited accuracy and efficiency of signal detection. Summary of the invention

[0007] The purpose of this application is to overcome the problems in the prior art of poor signal denoising and feature extraction and limited accuracy and efficiency of signal detection caused by the use of fixed wavelet basis functions, and to provide a TEOAE signal detection method and system based on adaptive multi-scale wavelet transform.

[0008] In a first aspect, a TEOAE signal detection method based on adaptive multi-scale wavelet transform is provided, comprising:

[0009] Collect TEOAE signals and perform preprocessing;

[0010] Extract key features from the preprocessed TEOAE signals;

[0011] Select the optimal wavelet basis function based on the extracted key features;

[0012] The preprocessed TEOAE signal is decomposed and synthesized at multiple scales to complete the wavelet transform;

[0013] The boundary effect elimination technology is used to eliminate the boundary effect of the signal after wavelet transformation;

[0014] Monitor detection results and adjust parameters in real time.

[0015] In some possible implementations, the preprocessing includes denoising and normalization, and the key features include mean, variance, and energy distribution.

[0016] In some possible implementations, the optimal wavelet basis function is selected according to the extracted key features, including:

[0017] Build a wavelet basis function library;

[0018] Calculating the correlation and energy distribution between each wavelet basis function in the wavelet basis function library and the preprocessed TEOAE signal;

[0019] The preprocessed TEOAE signal is subjected to wavelet transformation and reconstruction using each wavelet basis function, and the reconstruction error is calculated;

[0020] The optimal wavelet basis function is selected based on the comprehensive evaluation results of correlation, reconstruction error, energy distribution and complexity of wavelet basis function.

[0021] In some possible implementations, when selecting the optimal wavelet basis function, the weight ratio of the factors of correlation, reconstruction error, energy distribution, and complexity of the wavelet basis function is: 6:2:1:1.

[0022] In some possible implementations, a multi-scale wavelet transform is performed on the pre-processed TEOAE signal, including:

[0023] The preprocessed TEOAE signal is decomposed at multiple scales using the selected optimal wavelet basis function to generate a series of wavelet coefficients, where the decomposition formula is:

[0024]

[0025] in, is the preprocessed TEOAE signal, are the wavelet coefficients, is the wavelet basis function, is the decomposition error;

[0026] Dynamically adjust the decomposition scale according to the time-frequency characteristics of the preprocessed TEOAE signal;

[0027] The generated series of wavelet coefficients are synthesized at multiple scales to synthesize the final time domain signal, where the synthesis formula is:

[0028]

[0029] in, is the reconstructed signal, are the wavelet coefficients, is the wavelet basis function.

[0030] In some possible implementations, the decomposition scale mapping function in the multi-scale decomposition process of the pre-processed TEOAE signal is:

[0031]

[0032] in, is the decomposition scale at the current moment, is the mapping function, is the instantaneous frequency of the preprocessed TEOAE signal.

[0033] In some possible implementations, the boundary effect elimination technology includes zero filling technology, periodic extension, symmetric extension and interpolation technology, and a suitable boundary effect elimination technology is selected according to one or more factors of the boundary characteristics of the signal, analysis requirements and computing resources.

[0034] In some possible implementations, real-time monitoring of detection results and adjustment of parameters include:

[0035] Analyze the coefficients after wavelet transformation and evaluate the denoising effect and retention of signal characteristics;

[0036] Adjust the decomposition scale and wavelet basis function according to the analysis and evaluation results;

[0037] Record the adjusted decomposition scale and wavelet basis function.

[0038] In a second aspect, a TEOAE signal detection system based on adaptive multi-scale wavelet transform is provided, comprising:

[0039] Signal acquisition module, used to collect TEOAE signals and perform preprocessing;

[0040] Feature extraction module, used to extract key features from the preprocessed TEOAE signal;

[0041] A selection module, used to select the optimal wavelet basis function according to the extracted key features;

[0042] The transformation module is used to perform multi-scale decomposition and synthesis on the pre-processed TEOAE signal to complete the wavelet transformation;

[0043] A boundary effect elimination module is used to eliminate the boundary effect of the signal after wavelet transformation by using boundary effect elimination technology;

[0044] Feedback module, used to monitor the detection effect and adjust parameters in real time.

[0045] In a third aspect, a computer-readable storage medium is provided, wherein the computer-readable medium stores a program code for execution by a device, wherein the program code includes steps for executing the method in any one of the implementations of the first aspect described above.

[0046] The present application has the following beneficial effects: the present application can automatically select the optimal wavelet basis according to the signal characteristics, improve the accuracy and adaptability of signal detection, and can dynamically adjust the decomposition scale according to the instantaneous changes of the signal to better capture the transient characteristics of the signal. It can also optimize the detection process and improve the adaptability and robustness of the TEOAE signal detection method. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application.

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0049] Figure 1 is a flow chart of a TEOAE signal detection method based on adaptive multi-scale wavelet transform according to Example 1 of the present application;

[0050] Figure 2 is a flow chart of wavelet basis selection in the TEOAE signal detection method based on adaptive multi-scale wavelet transform in Example 1 of the present application;

[0051] Figure 3 It is a structural block diagram of a TEOAE signal detection system based on adaptive multi-scale wavelet transform according to Example 2 of the present application;

[0052] Reference numerals:

[0053] 100, signal acquisition module; 200, feature extraction module; 300, selection module; 400, transformation module; 500, boundary effect elimination module. DETAILED DESCRIPTION

[0054] 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.

[0055] Example 1

[0056] like Figure 1 As shown, a TEOAE signal detection method based on adaptive multi-scale wavelet transform according to Embodiment 1 of the present application includes:

[0057] S100, collecting TEOAE signals and performing preprocessing;

[0058] For example, to remove the noise of Teoae signal, you can select a conventional IIR filter, filter out the frequencies other than 1k-5k, and perform normalization.

[0059] It should be noted that the purpose of preprocessing the collected TEOAE signals is to improve the effect and accuracy of wavelet transform, which specifically includes the following points:

[0060] 1. Improve the signal-to-noise ratio: By removing the noise in the signal, the signal-to-noise ratio can be improved, so that the wavelet transform can capture the true characteristics of the signal more accurately.

[0061] 2. Remove irrelevant information: Preprocessing can remove irrelevant information or interference in the signal, such as trend terms, periodic interference, etc., so that the wavelet transform can focus more on the useful part of the signal.

[0062] 3. Enhance signal features: Through the preprocessing steps, specific features in the signal can be enhanced, such as edges, mutation points, etc. These features often correspond to important coefficients in the wavelet transform.

[0063] 4. Data normalization: Normalizing the signal can ensure that the wavelet transform of the signal at different scales is comparable, making it easier to select the appropriate wavelet basis and decomposition scale.

[0064] 5. Reduce computational complexity: By preprocessing and simplifying the signal, the amount of computation of wavelet transform can be reduced and computational efficiency can be improved.

[0065] 6. Meet the requirements of wavelet transform: Some wavelet transform algorithms require the signal to have a certain degree of stability or meet other conditions. Preprocessing can ensure that the signal meets these requirements.

[0066] 7. Improve the stability of wavelet coefficients: Preprocessing can reduce the instability of the signal, making the wavelet coefficients more stable, which is conducive to the selection of wavelet basis and scale.

[0067] 8. Easy to evaluate performance: The preprocessed signal is easier to evaluate because many interfering factors are removed, and the effects of different wavelet bases and scales on signal processing can be evaluated more accurately.

[0068] In this embodiment, through the above preprocessing steps, the adaptive wavelet basis selection mechanism can be made more effective, thereby improving the accuracy and robustness of the entire signal processing flow.

[0069] S200, extracting key features from the preprocessed TEOAE signal, such as mean, variance, energy distribution, etc.;

[0070] S300, select the optimal wavelet basis function according to the extracted key features, such as Figure 2 As shown, the specific steps include:

[0071] S301. Wavelet basis library construction: Create a wavelet basis function library, which contains a variety of wavelet bases (i.e. wavelet basis functions), including but not limited to Daubechies wavelet (dbN), Symlets wavelet (symN), Biorthogonal wavelet (biorNr.Nd), Coiflets wavelet (coifN), etc.

[0072] S302, correlation calculation: for each wavelet basis in the wavelet basis function library, calculate the correlation and energy distribution between it and the signal feature;

[0073] S303, reconstruction error evaluation: using each wavelet basis to perform wavelet transform and reconstruction on the signal, and calculate the reconstruction error;

[0074] S304, comprehensive evaluation: comprehensively evaluate each wavelet basis by combining correlation, reconstruction error, energy distribution and wavelet basis function complexity;

[0075] S305, selecting the optimal wavelet basis: selecting the optimal wavelet basis according to the comprehensive evaluation result;

[0076] The following is an introduction to combining correlation, reconstruction error, and energy distribution:

[0077] Correlation coefficient: used to measure the correlation between the wavelet basis and the signal;

[0078]

[0079] in, is the wavelet basis, It's a signal. is the mean of the wavelet basis, is the mean of the signal.

[0080] Mean square error: used to evaluate reconstruction error;

[0081]

[0082] in, is the mean square error, It's a signal. is the reconstruction signal.

[0083] Energy distribution matching: compare the energy distribution of the signal and the wavelet basis;

[0084]

[0085] in, is the energy distribution matching result, is the wavelet basis, It's a signal.

[0086] It should be noted that the basis for comprehensive evaluation includes the following factors:

[0087] Correlation: Select the wavelet basis that is most similar to the signal structure.

[0088] Reconstruction error: Select the wavelet basis that minimizes the reconstruction error.

[0089] Energy distribution: Select the wavelet basis whose energy distribution best matches the signal.

[0090] Complexity: Consider the complexity and computational efficiency of the wavelet basis.

[0091] In this embodiment, the weight ratio of the four factors of correlation, reconstruction error, energy distribution and complexity is 6:2:1:1.

[0092] In this embodiment, the optimal wavelet basis function is selected according to the comprehensive evaluation results of correlation, reconstruction error, energy distribution and wavelet basis function complexity, which improves the accuracy and adaptability of signal detection and reduces the complexity of artificially selecting wavelet basis.

[0093] S400, performing multi-scale decomposition and synthesis on the pre-processed TEOAE signal to complete wavelet transform;

[0094] It should be noted that multi-scale dynamic decomposition and synthesis is a signal processing technology that involves decomposing and synthesizing signals at different scales to capture the characteristics of signals in different time and frequency ranges. This technology is based on wavelet transform theory.

[0095] The specific steps include:

[0096] 1. Signal input: Use the wavelet basis selected by the adaptive wavelet basis mechanism for calculation.

[0097] 2. Multi-scale decomposition: Perform multi-scale decomposition on the signal to generate a series of wavelet coefficients.

[0098] 3. Dynamically adjust the decomposition scale: Dynamically adjust the decomposition scale according to the time-frequency characteristics of the signal.

[0099] 4. Wavelet synthesis: Perform multi-scale synthesis of the signal to synthesize the final time domain signal.

[0100] It should be noted that the dynamic adjustment of the decomposition scale is a key step in the multi-scale dynamic decomposition and synthesis technology, which allows the decomposition granularity to be automatically adjusted according to the time-frequency characteristics of the signal to better capture the characteristics of the signal in different time-frequency ranges. In the audio field, this usually means dynamically adjusting the decomposition scale according to the rhythm and melody of the music. The following is a simplified description of how to dynamically adjust the decomposition scale:

[0101] a. Instantaneous frequency estimation: Estimate the instantaneous frequency of the signal. The instantaneous frequency is the derivative of the instantaneous power spectrum of the signal, reflecting the instantaneous change of the signal in time.

[0102] b. Decomposition scale mapping function: define a mapping function to map the instantaneous frequency to the decomposition scale; the mapping function can be linear or nonlinear.

[0103] c. Dynamically adjust the decomposition scale: Based on the instantaneous frequency estimate, use a mapping function to calculate the current decomposition scale; update the decomposition scale to maintain the optimal analysis granularity in the dynamic changes of the signal.

[0104] d. Wavelet decomposition: Perform wavelet decomposition on the signal at a new decomposition scale to generate a series of wavelet coefficients.

[0105] The relevant calculation formula is as follows:

[0106] Wavelet decomposition:

[0107]

[0108] in, is the preprocessed TEOAE signal, are the wavelet coefficients, is the wavelet basis function, is the decomposition error.

[0109] Wavelet synthesis:

[0110]

[0111] in, is the reconstructed signal, are the wavelet coefficients, is the wavelet basis function.

[0112] Instantaneous frequency estimation:

[0113]

[0114] in, is the instantaneous frequency, is the Fourier transform, is a time domain signal.

[0115] Decomposition scale mapping function:

[0116]

[0117] Among them, is the decomposition scale at the current moment, is the mapping function, and is the instantaneous frequency of the preprocessed TEOAE signal.

[0118] The following is an example to further illustrate the decomposition scale:

[0119] In wavelet transform, the decomposition scale (also called decomposition level or number of layers) refers to the depth of wavelet decomposition of the signal. Each level of decomposition divides the signal into an approximate part (low-frequency part) and a detail part (high-frequency part). The decomposition scale determines the time-frequency resolution of the wavelet transform: a higher decomposition scale provides better frequency resolution but poorer time resolution, while a lower decomposition scale does the opposite.

[0120] Definition of decomposition scale: The decomposition scale is usually represented by an integer `j`, where `j=0` is the original signal, `j=1` is the first decomposition, and so on. In discrete wavelet transform (DWT), the decomposition scale `j` is related to the following factors: signal length `N`, wavelet filter length `L`, upsampling factor `2^j`.

[0121] Calculation of decomposition scale: The calculation of decomposition scale usually depends on the signal length and the length of the wavelet filter. For a given signal length `N` and filter length `L`, the maximum decomposition scale `J_max` can be calculated by the following formula:

[0122]

[0123] Here, `⌊·⌋` means rounding down. This formula ensures that after each layer of decomposition, the length of the approximation part is at least as long as the length of the wavelet filter, so that the next decomposition can be performed.

[0124] Dynamically adjusting the decomposition scale means that when analyzing a signal, the appropriate decomposition scale is automatically selected based on the signal's characteristics (such as frequency components, energy distribution, etc.). This can be achieved by following the steps below:

[0125] 1. Signal analysis: Analyze the characteristics of the signal and determine the frequency range that needs attention.

[0126] 2. Select the decomposition scale: According to the signal analysis results, select a suitable decomposition scale `j` so that the characteristics of the signal can be effectively represented at this scale.

[0127] 3. Implement wavelet transform: Perform wavelet transform on the signal at the selected scale `j`.

[0128] Example: Assuming there is a signal of length `N=256` and the wavelet filter length used is `L=10` (this is an assumed number, the actual wavelet filter length depends on the specific wavelet basis), then the maximum decomposition scale `J_max` can be calculated as follows:

[0129]

[0130] J_max=4, which means that the signal can be decomposed by up to 4 layers of wavelet.

[0131] In this way, the system can automatically adjust the decomposition scale to adapt to the instantaneous changes of the signal. This ability to dynamically adjust the decomposition scale makes the multi-scale dynamic decomposition and synthesis technology particularly effective when processing non-stationary signals. In practical applications, the mapping function can be optimized according to the specific characteristics of the signal to obtain the best analysis effect.

[0132] S500, using boundary effect elimination technology to eliminate the boundary effect of the signal after wavelet transformation;

[0133] It should be noted that the boundary effect elimination technology is an improved signal processing method that aims to solve the distortion problem caused by signal boundary truncation in wavelet transform. This technology effectively reduces the boundary effect and improves the accuracy and reliability of signal processing by introducing a specific boundary processing strategy.

[0134] Boundary effect elimination techniques include the following:

[0135] 1. Zero-Padding:

[0136] Add zero values ​​at the boundaries of a signal to make the length of the signal an integer multiple of its original length.

[0137] Then the wavelet transform is performed, and the value of the signal at the boundary will not affect the analysis of other parts.

[0138] Finally, the added zero values ​​are removed and the signal is reconstructed.

[0139] 2. Periodic Extension:

[0140] The signal is considered as a periodic signal and the wavelet transform is performed using periodic boundary conditions.

[0141] Periodic extension makes the signal value at the boundary periodic, thus eliminating the boundary effect.

[0142] 3. Symmetric Extension:

[0143] Adds values ​​symmetrical to the signal at its boundaries.

[0144] Then the wavelet transform is performed, and the value of the signal at the boundary will not affect the analysis of other parts.

[0145] Finally, the added values ​​are removed and the signal is reconstructed.

[0146] 4. Interpolation technology:

[0147] Use interpolation methods (such as linear interpolation, spline interpolation, etc.) to estimate the value of the signal at the boundary.

[0148] Then the wavelet transform is performed, and the value of the signal at the boundary will not affect the analysis of other parts.

[0149] Finally, remove the interpolated values ​​and reconstruct the signal.

[0150] The basis for selecting boundary effect elimination technology:

[0151] 1. Signal characteristics: Select the appropriate boundary processing strategy based on the characteristics of the signal.

[0152] If the signal is periodic at the boundaries, period continuation can be used.

[0153] If the signal is symmetric at the boundary, you can use symmetric continuation.

[0154] If the signal is stationary at the boundaries, zero padding can be used.

[0155] If the signal is non-stationary at the boundaries, interpolation techniques can be used.

[0156] 2. Analyze requirements: Select appropriate boundary processing strategies based on analysis requirements.

[0157] If the analysis needs to preserve the local characteristics of the signal, zero padding or symmetric extension should be selected.

[0158] If the analysis needs to preserve the overall characteristics of the signal, periodic extension should be selected.

[0159] If the analysis requires preserving the continuous characteristics of the signal, an interpolation technique should be chosen.

[0160] 3. Computing resources: Consider computing resource limitations and choose appropriate boundary processing strategies.

[0161] Zero padding and cycle extension usually require less computational resources.

[0162] Symmetric continuation and interpolation techniques may require more computing resources.

[0163] According to the characteristics of TEOAE signals, the signals are non-stationary at the boundaries and are filled using interpolation techniques.

[0164] It should be noted that interpolation technology is a mathematical method used to estimate the value of a function at an unknown point. In signal processing, especially when dealing with boundary effects in wavelet transform, interpolation technology can be used to fill the boundaries of the signal to reduce boundary effects. The following is a common interpolation method - linear interpolation, as well as its calculation method and steps:

[0165] Linear interpolation: Linear interpolation is the simplest interpolation method, which assumes that the function between two known points is linear.

[0166] Calculation method of linear interpolation:

[0167] Assuming there are two known points (x0, y0) and (x1, y1), and the point to be interpolated is x, then the linear interpolation formula is:

[0168]

[0169] A specific example will be used to demonstrate how to use linear interpolation technology to fill the signal boundary (the length of this example may be inappropriate, generally the TEOAE signal is longer, this can only be used as an example to show how to calculate). Assuming there is a short signal and using the RBI03.3 wavelet basis. Linear interpolation will be used to fill the boundary.

[0170] Step 1: Define the original signal

[0171] Assume the original signal is:

[0172]

[0173] Step 2: Determine the filter length

[0174] This is related to the decomposition scale of the wavelet filter. Assume that the length of the low-pass and high-pass filters of the rbio3.3 wavelet basis is 6. Therefore, 5 points need to be filled at both ends of the signal (6-1=5).

[0175] Step 3: Linear interpolation filling

[0176] Linear interpolation is required at both ends of the signal. The following are the steps for linear interpolation:

[0177] Left border padding:

[0178] Use the first point in the signal and the second point Perform linear interpolation. The interpolation formula is: For the left border, fill in 5 points from arrive .

[0179] Calculate the first interpolation point :

[0180]

[0181] Calculate the second interpolation point :

[0182]

[0183] Similarly, other points can be calculated:

[0184]

[0185]

[0186] It is worth noting that due to It is already part of the original signal and does not need to be recalculated.

[0187] Right fill: Use the last point in the signal and the second to last point For the right border, fill in 5 points from arrive .

[0188] Calculate the first interpolation point :

[0189]

[0190] Calculate the second interpolation point :

[0191]

[0192] Similarly, other points can be calculated:

[0193]

[0194]

[0195]

[0196] Step 4: Combine the filled signals

[0197] Combine the left boundary, the original signal, and the right boundary to get the padded signal:

[0198]

[0199] In this way, the filling of the signal boundaries using linear interpolation techniques is completed. This filled signal can now be used for wavelet transform without being affected by boundary effects.

[0200] It is worth noting that linear interpolation is only one of many interpolation methods. In practical applications, it may be necessary to select a more appropriate interpolation method based on the characteristics of the signal, such as polynomial interpolation, spline interpolation, etc.

[0201] In this embodiment, the boundary effect of the signal after wavelet transformation is eliminated by boundary effect elimination technology, which effectively reduces the boundary effect and improves the accuracy and reliability of signal processing. At the same time, it can reduce the distortion at the boundary and improve the stability of the signal processing results. In addition, a variety of boundary processing strategies are provided, which can be selected and optimized according to specific needs, so that the present application can be applicable to various types of TEOAE signals, including non-stationary signals and transient signals.

[0202] S600, monitor the detection effect in real time and adjust the parameters. Specifically, dynamically optimize the decomposition scale and wavelet basis of the wavelet transform according to the signal-to-noise ratio (SNR). The following is a description of the optimization process and its logical steps:

[0203] Step 1: Real-time signal monitoring

[0204] The system continuously monitors the input signal to detect changes in its characteristics, such as amplitude, frequency content, etc.

[0205] Step 2: Calculate the real-time signal-to-noise ratio

[0206] To calculate the signal-to-noise ratio of a signal in real time, the following formula can be used:

[0207]

[0208] in, is the signal power, is the noise power.

[0209] Step 3: Determine the signal-to-noise ratio range

[0210] If SNR > 10 dB, the signal quality is considered good.

[0211] If 0 dB < SNR ≤ 10 dB, the signal is considered to contain some noise.

[0212] If SNR ≤ 0 dB, the signal-to-noise level is considered high.

[0213] Step 4: Select the optimal decomposition scale

[0214] For high signal-to-noise ratio signals, a smaller resolution scale is chosen to preserve signal details.

[0215] For signals with medium signal-to-noise ratio, a medium decomposition scale is selected to balance signal feature preservation and noise removal.

[0216] For signals with low signal-to-noise ratio, a larger decomposition scale is selected to better remove noise.

[0217] Step 5: Select the optimized wavelet basis

[0218] According to the signal characteristics and signal-to-noise ratio, select the appropriate wavelet basis:

[0219] High signal-to-noise ratio: Select a wavelet basis with higher regularity, such as the coif series.

[0220] Medium signal-to-noise ratio: Choose a wavelet basis that balances denoising and feature preservation, such as the DB or SYM series.

[0221] Low signal-to-noise ratio: Select a wavelet basis with strong denoising ability, such as the Bior series.

[0222] Step 6: Implement wavelet transform

[0223] Perform wavelet transform on the signal using the selected decomposition scale and wavelet basis.

[0224] Step 7: Evaluate the transformation

[0225] Analyze the transformed coefficients to evaluate the denoising effect and the retention of signal characteristics.

[0226] Step 8: Adjust parameters (if necessary)

[0227] If the evaluation results show that the signal quality does not meet the requirements, the parameters are adjusted according to the following logic:

[0228] If the noise level is still high, increase the decomposition scale or replace the wavelet basis with stronger denoising ability;

[0229] If the signal features are lost, reduce the decomposition scale or replace the wavelet basis with a higher regularity.

[0230] Step 9: Record and Feedback

[0231] The optimized parameter settings are recorded and the optimization results are fed back to the system for the next round of real-time detection and adjustment.

[0232] Through the above logical steps, the decomposition scale and the selection of wavelet basis can be optimized in real time according to the signal-to-noise ratio of the signal, thereby maintaining the efficiency and accuracy of signal processing in a dynamically changing environment.

[0233] It should be noted that the present application provides unique advantages by combining the adaptability and dynamic adjustment characteristics of wavelet transform, especially in processing non-stationary signals and transient feature extraction. The present application can automatically select the optimal wavelet basis according to the signal characteristics, thereby improving the accuracy and adaptability of signal detection, and can dynamically adjust the decomposition scale according to the instantaneous changes of the signal to better capture the transient characteristics of the signal. It can also optimize the detection process and improve the adaptability and robustness of the TEOAE signal detection method.

[0234] Example 2

[0235] like Figure 3 As shown, a TEOAE signal detection system based on adaptive multi-scale wavelet transform according to Embodiment 2 of the present application includes:

[0236] A signal acquisition module 100 is used to acquire TEOAE signals and perform preprocessing;

[0237] A feature extraction module 200 is used to extract key features from the preprocessed TEOAE signal;

[0238] A selection module 300 is used to select an optimal wavelet basis function according to the extracted key features;

[0239] Transformation module 400, used for performing multi-scale decomposition and synthesis on the pre-processed TEOAE signal to complete wavelet transformation;

[0240] The boundary effect elimination module 500 is used to eliminate the boundary effect of the signal after wavelet transformation by using the boundary effect elimination technology;

[0241] The feedback module 600 is used to monitor the detection effect and adjust the parameters in real time.

[0242] It should be noted that other specific implementations of the TEOAE signal detection system based on adaptive multi-scale wavelet transform in this embodiment can refer to the specific implementations of the TEOAE signal detection method based on adaptive multi-scale wavelet transform mentioned above, which will not be repeated here to avoid redundancy.

[0243] Example 3

[0244] A computer-readable storage medium according to Embodiment 3 of the present application, wherein the computer-readable storage medium stores a program code for execution by a device, wherein the program code includes steps for executing a method in any one of the implementations in Embodiment 1 of the present application;

[0245] Among them, the computer-readable storage medium can be a read-only memory (ROM), a static storage device, a dynamic storage device or a random access memory (RAM); the computer-readable storage medium can store program code, and when the program stored in the computer-readable storage medium is executed by the processor, the processor is used to execute the steps of the method in any one of the implementation methods in Example 1 of the present application.

[0246] The above are only preferred specific implementations of the present application; however, the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical solution and its improved ideas of the present application within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A TEOAE signal detection method based on adaptive multi-scale wavelet transform, characterized in that: include: Collect TEOAE signals and perform preprocessing; Extract key features from the preprocessed TEOAE signals; Select the optimal wavelet basis function based on the extracted key features; The preprocessed TEOAE signal is decomposed and synthesized at multiple scales to complete the wavelet transform; The boundary effect elimination technology is used to eliminate the boundary effect of the signal after wavelet transformation; Monitor detection results and adjust parameters in real time.

2. The TEOAE signal detection method based on adaptive multi-scale wavelet transform according to claim 1 is characterized in that: The preprocessing includes denoising and normalization, and the key features include mean, variance and energy distribution.

3. The TEOAE signal detection method based on adaptive multi-scale wavelet transform according to claim 1 is characterized in that: The optimal wavelet basis function is selected based on the extracted key features, including: Build a wavelet basis function library; Calculating the correlation and energy distribution between each wavelet basis function in the wavelet basis function library and the preprocessed TEOAE signal; The preprocessed TEOAE signal is subjected to wavelet transformation and reconstruction using each wavelet basis function, and the reconstruction error is calculated; The optimal wavelet basis function is selected based on the comprehensive evaluation results of correlation, reconstruction error, energy distribution and complexity of wavelet basis function.

4. The TEOAE signal detection method based on adaptive multi-scale wavelet transform according to claim 3 is characterized in that: When selecting the optimal wavelet basis function, the weight ratio of correlation, reconstruction error, energy distribution, and wavelet basis function complexity is 6:2:1:

1.

5. The TEOAE signal detection method based on adaptive multi-scale wavelet transform according to claim 1 is characterized in that: Perform multi-scale wavelet transform on the pre-processed TEOAE signal, including: The preprocessed TEOAE signal is decomposed at multiple scales using the selected optimal wavelet basis function to generate a series of wavelet coefficients, where the decomposition formula is: in, is the preprocessed TEOAE signal, are the wavelet coefficients, is the wavelet basis function, is the decomposition error; Dynamically adjust the decomposition scale according to the time-frequency characteristics of the preprocessed TEOAE signal; The generated series of wavelet coefficients are synthesized at multiple scales to synthesize the final time domain signal, where the synthesis formula is: in, is the reconstructed signal, are the wavelet coefficients, is the wavelet basis function.

6. The TEOAE signal detection method based on adaptive multi-scale wavelet transform according to claim 5 is characterized in that: The decomposition scale mapping function in the multi-scale decomposition process of the preprocessed TEOAE signal is: in, is the decomposition scale at the current moment, is the mapping function, is the instantaneous frequency of the preprocessed TEOAE signal.

7. The TEOAE signal detection method based on adaptive multi-scale wavelet transform according to claim 1 is characterized in that: The boundary effect elimination technology includes zero filling technology, periodic extension, symmetric extension and interpolation technology. The appropriate boundary effect elimination technology is selected according to one or more factors of the boundary characteristics of the signal, analysis requirements and computing resources.

8. The TEOAE signal detection method based on adaptive multi-scale wavelet transform according to claim 1 is characterized in that: Monitor detection results in real time and adjust parameters, including: Analyze the coefficients after wavelet transformation and evaluate the denoising effect and retention of signal characteristics; Adjust the decomposition scale and wavelet basis function according to the analysis and evaluation results; Record the adjusted decomposition scale and wavelet basis function.

9. A TEOAE signal detection system based on adaptive multi-scale wavelet transform, characterized in that: include: Signal acquisition module, used to collect TEOAE signals and perform preprocessing; Feature extraction module, used to extract key features from the preprocessed TEOAE signal; A selection module, used to select the optimal wavelet basis function according to the extracted key features; The transformation module is used to perform multi-scale decomposition and synthesis on the pre-processed TEOAE signal to complete the wavelet transformation; A boundary effect elimination module is used to eliminate the boundary effect of the signal after wavelet transformation by using boundary effect elimination technology; Feedback module, used to monitor the detection effect and adjust parameters in real time.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program code for execution by a device, wherein the program code includes steps for executing the method according to any one of claims 1 to 8.

Images