An optimized otoacoustic emission signal detection method

Abstract

The application provides an optimized ear sound reflection signal detection method, aiming at improving the accuracy of quality analysis and abnormal identification of ear sound signals, and mainly comprises the following steps: TE noise threshold optimization, artifact elimination, similarity calculation, wavelet denoising and DP data processing, etc., wherein the self-adaptive threshold control is used to dynamically adjust the noise threshold, thereby enhancing the reliability of the signal; the T-shaped window and the band-pass filtering technology are used to effectively filter the noise and improve the signal-to-noise ratio; the frequency domain coherence spectrum method and the wavelet denoising further strengthen the signal matching degree and stability; and the continuous sound processing and the small cavity identification strategy ensure the data integrity and accurately distinguish the influence of different environments and earplug states.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of objective hearing detection based on ear acoustics, in particular to an optimized ear acoustic reflex signal detection method. BACKGROUND

[0002] OAE (auditory evoked response) signals are audio signals generated spontaneously from the ear canal, which are usually closely related to hearing function. Therefore, accurately detecting and analyzing OAE signals is of great significance for evaluating hearing health status. However, due to environmental noise and equipment limitations, traditional OAE signal detection methods face many challenges in practical applications. Existing OAE signal detection methods mainly rely on the performance of filters, and different types and orders of filters will result in different results. In addition, under low signal-to-noise ratio conditions, accurately identifying and quantifying the linear relationship between the two signals is also a big problem. These problems limit the accuracy and reliability of OAE signal detection. Therefore, we propose an optimized ear acoustic reflex signal detection method. SUMMARY

[0003] The purpose of the present application is to provide an optimized ear acoustic reflex signal detection method to solve the problems raised in the background.

[0004] To achieve the above purpose, the present application provides the following technical scheme: an optimized ear acoustic reflex signal detection method, comprising the following steps:

[0005] S11, noise threshold rejection, through a self-defined control program calculation process, the detection accuracy and the detection time are balanced, and the parameter setting includes threshold upper and lower limit, peak characteristic number, similarity weight ratio column and the number of conforming weight judgment standard parameter setting, through signal preprocessing, the better and reliable data block is reserved, and the next calculation is carried out;

[0006] S12, artifact elimination includes 4.2ms filter, data length padding and band pass filter design, through the combination of T type window data processing method and filtering technology, the coherent average method calculation is carried out, the linear artifact is filtered out, and the signal-to-noise ratio of TEOAE is improved;

[0007] S13, similarity calculation includes frequency domain conversion calculation, frequency domain spectrum calculation, group signal similarity calculation, and signal matching through coherent spectrum method, which enhances the quantitative evaluation of signal linearity under low signal-to-noise ratio conditions;

[0008] S14, through wavelet decomposition and wavelet reconstruction, the reference signal with lower noise is obtained, according to the signal property, the key frequency point of the signal is extracted, which is used for TEOAE signal-to-noise ratio calculation.

[0009] S15. Through data cycle truncation, Fourier transform and audio breakpoint removal, continuous sound processing is obtained to reduce noise interference and highlight signal characteristics. Then, the frequency amplitude at the third-order intermodulation point 2f1-f2 is detected and calculated, and the average amplitude of the three points around 2f1-f2 is calculated to obtain the signal-to-noise ratio of DPOAE.

[0010] S16, generation of sweep frequency signal and sweep frequency detection threshold, accurately identify small cavity types and improve small cavity identification accuracy.

[0011] Preferably, the S11 noise threshold removal algorithm includes the following steps:

[0012] S21, using a T-shaped window as a window function, performing window function calculation on 1253 data lengths;

[0013] S22, filter out the stimulus signal 4.2 milliseconds before the signal, and fill in the signal between 1254 and 2048 milliseconds;

[0014] S23, performing a band-pass filtering method to remove noise interference outside the 500-6000 Hz range;

[0015] S24. Calculate the coherent average of the signal to remove linear artifacts and improve the signal-to-noise ratio of TEOAE.

[0016] Preferably, the S12 is implemented by an artifact elimination algorithm to remove linear artifacts of the sample.

[0017] Preferably, the step S13 is implemented by similarity calculation, using a frequency domain coherence spectrum method to calculate the similarity between the two signals, and identifying and quantifying the linear relationship between the two signals.

[0018] Preferably, the S14 applies a specific decomposition order to the stationary and non-stationary signals through wavelet denoising to eliminate noise signals.

[0019] Preferably, the wavelet denoising and similarity calculation includes the following steps:

[0020] S31, performing third-order wavelet decomposition and reconstruction on the audio data to obtain a noise-reduced signal;

[0021] S32, performing frequency domain conversion on the reconstructed signal;

[0022] S33, calculating the coherence spectra of the two groups of signals in the frequency domain;

[0023] S34. Calculate the similarity between the two signals according to the coherence spectrum result.

[0024] Preferably, the S15 is implemented through DP continuous sound data processing and characteristic signal detection algorithm, removing each breakpoint to make the audio into continuous sound, thereby reducing the overall increase in background noise caused by the breakpoint and improving the accuracy and reliability of DP data processing.

[0025] Preferably, the DP detection algorithm comprises the following steps:

[0026] S41. Removing one distortion point signal every 1025 data lengths can reduce the power spectrum noise floor, highlight the frequency amplitude of the third-order signal, and improve the signal-to-noise ratio;

[0027] S42. Detect the frequency amplitude at 2f1-f2 and calculate the signal strength.

[0028] S43. Calculate the average amplitude of the three points around 2f1-f2 to obtain the signal-to-noise ratio.

[0029] Preferably, the S16 is implemented by a small cavity recognition algorithm, which calculates the smoothness and energy distribution range of the data based on the integrity of the single frequency sweep test data, and distinguishes and recognizes the small cavity type and the state of the earplug.

[0030] Preferably, the small cavity identification algorithm comprises the following steps:

[0031] S51, step 1: generating a sweep frequency signal of 20-6000 (Hz) and emitting sound;

[0032] S52. Calculate the smoothness and energy distribution range of the received data according to the distribution of the frequency sweep signal;

[0033] S53. Setting threshold ranges for different types of small cavities based on the data distribution range.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] (1) Through the custom control program calculation process, balance the detection accuracy and detection time.

[0036] (2) First, the signal is filtered using the set T-shaped window as the window function.

[0037] (3) This method uses frequency domain coherence spectrum to calculate the similarity between two signals. It can effectively identify and quantify the linear relationship between the two signals and can calculate the similarity even when the signal-to-noise ratio is low. In addition, frequency domain coherence spectrum can more effectively distinguish between true mutual relationships and accidental, meaningless correlations.

[0038] (4) Performing frequency domain transformation on the denoised audio data reveals that signal peaks can be found in certain frequency bands. These peaks can be used as five signal frequency points for signal-to-noise ratio calculation.

[0039] (5) Removing breakpoints ensures the continuity of signal data, reduces audio background noise, improves signal reliability, enables better calculation of signal-to-noise ratio, and improves the efficiency of otoacoustic detection.

[0040] (6) The energy distribution interval of the signal sound pressure level of the sweep frequency detection is clearly distinguished, and a reasonable threshold range can be divided to eliminate the interference of small cavities and the third-order intermodulation interference in some spaces. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a flowchart of the implementation method of the present invention;

[0042] Figure 2 Implementation case test result graph for DPOAE;

[0043] Figure 3 Figure 2. Test results for the TEOAE implementation case. DETAILED DESCRIPTION

[0044] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0045] See also Figure 1-3 The present invention provides a technical solution: an optimized otoacoustic reflex signal detection method, comprising the following steps:

[0046] S11, noise threshold removal, through the custom control program calculation process, weighing the detection accuracy and detection time, setting parameters including the upper and lower threshold limits, peak feature number, similarity weight ratio and the number of weighted judgment criteria, through signal preprocessing, retaining the better and more reliable data blocks for the next step of calculation;

[0047] S12, artifact elimination includes filtering out the first 4.2ms, data length padding, and bandpass filtering. By combining T-window data processing with filtering technology, coherent averaging is performed to filter out linear artifacts and improve the signal-to-noise ratio of TEOAE.

[0048] S13. Similarity calculation includes frequency domain conversion calculation, frequency domain spectrum coherence calculation, group signal similarity calculation, signal matching through coherence spectrum method, and enhanced quantitative evaluation of signal linearity under low signal-to-noise ratio conditions;

[0049] S14. Through wavelet decomposition and wavelet reconstruction, a reference signal with low background noise is obtained. According to the signal properties, the key frequency points of the signal are extracted for TEOAE signal-to-noise ratio calculation.

[0050] S15. Through data cycle truncation, Fourier transform and audio breakpoint removal, continuous sound processing is obtained to reduce noise interference and highlight signal characteristics. Then, the frequency amplitude at the third-order intermodulation point 2f1-f2 is detected and calculated, and the average amplitude of the three points around 2f1-f2 is calculated to obtain the signal-to-noise ratio of DPOAE.

[0051] S16, generation of sweep frequency signal and sweep frequency detection threshold, accurately identify small cavity types and improve small cavity identification accuracy.

[0052] As a preferred embodiment, the S11 noise threshold removal algorithm includes the following steps:

[0053] S21, using a T-shaped window as a window function, performing window function calculation on 1253 data lengths;

[0054] S22, filter out the stimulus signal 4.2 milliseconds before the signal, and fill in the signal between 1254 and 2048 milliseconds;

[0055] S23, performing a band-pass filtering method to remove noise interference outside the 500-6000 Hz range;

[0056] S24. Calculate the coherent average of the signal to remove linear artifacts and improve the signal-to-noise ratio of TEOAE.

[0057] As a preferred implementation, the step S12 is implemented by an artifact elimination algorithm to remove linear artifacts of the sample.

[0058] As a preferred implementation, S13 is implemented by similarity calculation, using a frequency domain coherence spectrum method to calculate the similarity between the two signals, and identifying and quantifying the linear relationship between the two signals.

[0059] As a preferred implementation, the S14 applies a specific decomposition order to the stationary and non-stationary signals through wavelet denoising to eliminate noise signals.

[0060] As a preferred embodiment, wavelet denoising and similarity calculation include the following steps:

[0061] S31, performing third-order wavelet decomposition and reconstruction on the audio data to obtain a noise-reduced signal;

[0062] S32, performing frequency domain conversion on the reconstructed signal;

[0063] S33, calculating the coherence spectra of the two groups of signals in the frequency domain;

[0064] S34, calculating the similarity between the two signals according to the result of the coherent spectrum.

[0065] As a preferred embodiment, the S15 is realized by a DP continuous tone data processing and feature signal detection algorithm, each breakpoint is removed so that the audio becomes continuous tone, thereby reducing the overall noise floor caused by the breakpoint, and improving the accuracy and reliability of the DP data processing.

[0066] As a preferred embodiment, the DP detection algorithm comprises the following steps:

[0067] S41, removing one distortion point signal every 1025 data lengths can reduce the power spectrum noise floor, highlight the frequency amplitude of the third-order signal, and improve the signal-to-noise ratio;

[0068] S42, detecting the frequency amplitude at 2f1-f2, and calculating the strength of the signal.

[0069] S43, calculating the average amplitude of the three points around 2f1-f2 to obtain the signal-to-noise ratio.

[0070] As a preferred embodiment, the S16 is realized by a small cavity recognition algorithm, which calculates the smoothness and energy distribution interval of the single sweep frequency test data according to the integrity of the data, and distinguishes and recognizes the small cavity type and the state of the earplug.

[0071] As a preferred embodiment, the small cavity recognition algorithm comprises the following steps:

[0072] S51, first, generate a sweep frequency signal of 20-6000 (Hz) and sound;

[0073] S52, according to the distribution of the sweep frequency signal, calculate the smoothness and energy distribution interval of the received data;

[0074] S53, according to the distribution interval of the data, set the threshold range of different types of small cavities.

[0075] The working principle of the application is: TEOAE algorithm optimization, first, collect data and dynamically set appropriate signal threshold upper and lower limits, cycle sample value and smooth step length, calculate and identify the peak feature number exceeding the noise threshold, effectively distinguish useful signals and noise signals, and retain important audio data; then apply a T window function to initialize the signal processing.

[0076] The first 4.2ms of data were eliminated to prevent the stimulation signal from affecting the results. Afterwards, data of different lengths were padded, and bandpass filtering was used to remove signals outside the specific frequency range. The coherent averaging method was used to filter out linear artifacts and improve the signal-to-noise ratio of the TEOAE test. Finally, the similarity and signal-to-noise ratio of the data after wavelet denoising were calculated, and the detection data were displayed.

[0077] DPOAE time-domain data processing first generates a specific frequency sweep signal based on a predetermined threshold, and calculates the power spectral density and peak energy output of the echo signal to distinguish different test environments. If the set threshold is met, the original sound data obtained is truncated periodically and Fourier transformed. Then, an algorithm is used to identify and remove audio breakpoints, and the data is converted into continuous sound. This process reduces the noise interference introduced by the breakpoints and clearly highlights the signal characteristics. Finally, combined with the frequency amplitude calculation of the third-order intermodulation point, the key frequency points such as 2f1-f2 are analyzed, and the sound pressure levels of the signal and noise are calculated to evaluate the signal-to-noise ratio of DPOAE.

[0078] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0079] While 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 these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An optimized otoacoustic reflex signal detection method, characterized in that: The following steps are involved: S11, noise threshold removal, through the custom control program calculation process, weighing the detection accuracy and detection time, setting parameters including the upper and lower threshold limits, peak feature number, similarity weight ratio and the number of weighted judgment criteria, through signal preprocessing, retaining the better and more reliable data blocks for the next step of calculation; S12, artifact elimination includes filtering out the first 4.2ms, data length padding, and bandpass filtering. By combining T-window data processing with filtering technology, coherent averaging is performed to filter out linear artifacts and improve the signal-to-noise ratio of TEOAE. S13. Similarity calculation includes frequency domain conversion calculation, frequency domain spectrum coherence calculation, group signal similarity calculation, signal matching through coherence spectrum method, and enhanced quantitative evaluation of signal linearity under low signal-to-noise ratio conditions; S14. Obtain a reference signal with low background noise through wavelet decomposition and wavelet reconstruction. Extract key frequency points of the signal according to the signal properties for use in TEOAE signal-to-noise ratio calculation. S15. Continuous sound processing is performed by data cycle truncation, Fourier transform, and audio breakpoint removal to reduce noise interference and highlight signal characteristics. The frequency amplitude at the third-order intermodulation point 2f1-f2 is then detected and calculated. The average amplitude of the three points around 2f1-f2 is calculated to obtain the signal-to-noise ratio of the DPOAE. S16, generation of sweep frequency signal and sweep frequency detection threshold, accurately identify small cavity types and improve small cavity identification accuracy.

2. The optimized otoacoustic reflex signal detection method according to claim 1, wherein: The S11 noise threshold rejection algorithm includes the following steps: S21, using a T-shaped window as a window function, performing window function calculation on 1253 data lengths; S22, filter out the stimulus signal 4.2 milliseconds before the signal, and fill in the signal between 1254 and 2048 milliseconds; S23, performing a band-pass filtering method to remove noise interference outside the 500-6000 Hz range; S24. Calculate the coherent average of the signal to remove linear artifacts and improve the signal-to-noise ratio of TEOAE.

3. The optimized otoacoustic reflex signal detection method according to claim 1, wherein: The S12 is implemented by an artifact elimination algorithm to remove linear artifacts of the sample.

4. The optimized otoacoustic reflex signal detection method according to claim 1, wherein: The step S13 is implemented by similarity calculation, using a frequency domain coherence spectrum method to calculate the similarity between the two signals and identify and quantify the linear relationship between the two signals.

5. The optimized otoacoustic reflex signal detection method according to claim 4, wherein: The S14 applies a specific decomposition order to the stationary and non-stationary signals through wavelet denoising to eliminate noise signals.

6. The optimized otoacoustic reflex signal detection method according to claim 5, wherein: Wavelet denoising and similarity calculation include the following steps: S31, performing third-order wavelet decomposition and reconstruction on the audio data to obtain a noise-reduced signal; S32, performing frequency domain conversion on the reconstructed signal; S33, calculating the coherence spectra of the two groups of signals in the frequency domain; S34. Calculate the similarity between the two signals according to the coherence spectrum result.

7. The optimized otoacoustic reflex signal detection method according to claim 1, wherein: The S15 is implemented through DP continuous sound data processing and characteristic signal detection algorithm, removing each breakpoint to make the audio continuous sound, thereby reducing the overall increase in background noise caused by the breakpoint and improving the accuracy and reliability of DP data processing.

8. The optimized otoacoustic reflex signal detection method according to claim 7, wherein: The DP detection algorithm consists of the following steps: S41. Removing one distortion point signal every 1025 data lengths can reduce the power spectrum noise floor, highlight the frequency amplitude of the third-order signal, and improve the signal-to-noise ratio; S42, detecting the frequency amplitude at 2f1-f2 and calculating the signal strength; S43. Calculate the average amplitude of the three points around 2f1-f2 to obtain the signal-to-noise ratio.

9. The optimized otoacoustic reflex signal detection method according to claim 1, wherein: The S16 is implemented by a small cavity recognition algorithm, which calculates the smoothness and energy distribution range of the data based on the integrity of the single frequency sweep test data, and distinguishes and recognizes the small cavity type and the status of the earplug.

10. The optimized otoacoustic reflex signal detection method according to claim 9, characterized in that: The small cavity recognition algorithm includes the following steps: S51, step 1: generate a sweep frequency signal of 20-6000 Hz and make a sound; S52. Calculate the smoothness and energy distribution range of the received data according to the distribution of the frequency sweep signal; S53. Setting threshold ranges for different types of small cavities based on the data distribution range.

Images