Wideband adaptive filtering and inverse sound wave synthesis method for intelligent noise reduction device

Abstract

This invention relates to the field of sound processing technology, specifically to a broadband adaptive filtering and inverse acoustic synthesis method for intelligent noise reduction equipment. The method includes: performing local mean decomposition on the original audio signal to obtain multiple product function components and a monotonic residual signal; analyzing each product function component based on the original audio signal to obtain a device characteristic index for each product function component; analyzing the amplitude variation of each product function component to obtain a human voice interference index for each product function component; determining the interference weight of each product function component based on its device characteristic index and human voice interference index; and obtaining the corresponding noise interference signal based on the multiple product function components, the monotonic residual signal, and the interference weight of each product function component. This invention can accurately identify environmental noise and human voice interference during equipment operation, thereby improving the noise reduction effect of industrial audio.

Description

Technical Field

[0001] This invention relates to the field of sound processing technology, specifically to a broadband adaptive filtering and reverse sound wave synthesis method for intelligent noise reduction devices. Background Technology

[0002] In voice input scenarios such as voice communication, speech recognition, and human-computer interaction, background environmental noise can significantly impact the clarity and recognizability of speech signals. This is especially true in open spaces or complex noisy environments, where noise frequencies are wide, components are complex, and changes rapidly. The use of intelligent noise reduction devices can effectively suppress noise interference in speech.

[0003] Intelligent noise reduction devices are typically equipped with active noise cancellation (ANC) and have the ability to automatically identify noise characteristics and dynamically adjust noise reduction strategies. The ANC function mainly neutralizes noise by synthesizing inverse sound waves, thereby achieving the noise reduction effect.

[0004] Current active noise cancellation technology is mainly applied in enclosed environments, primarily targeting low-frequency narrowband noise, and is difficult to apply to open or wideband multi-source noise environments. Meanwhile, traditional adaptive filtering techniques are mostly fixed-bandwidth or single-frequency filtering, lacking the dynamic adaptability to complex wideband noise environments. Especially in industrial acoustic condition monitoring, it is difficult to analyze and eliminate noise that overlaps with the operating frequency bands of some equipment, easily affecting the sound quality of the target audio and failing to achieve relatively ideal noise reduction results.

[0005] In other words, existing active noise cancellation solutions are not very effective at reducing noise in industrial audio. Summary of the Invention

[0006] To address the problem of poor noise reduction performance of existing active noise cancellation solutions for industrial audio, this invention aims to provide a broadband adaptive filtering and inverse acoustic synthesis method for intelligent noise cancellation devices. The specific technical solution adopted is as follows:

[0007] In a first aspect, one embodiment of the present invention provides a broadband adaptive filtering and inverse acoustic wave synthesis method for an intelligent noise reduction device, the method comprising:

[0008] The original audio signal is subjected to local mean decomposition to obtain multiple product function components and a monotonic residual signal. The original audio signal is derived from an industrial production environment.

[0009] Based on the analysis of each product function component of the original audio signal, the device characteristic index of each product function component is obtained. The device characteristic index is used to represent the probability of the product function component corresponding to the sound of equipment operation in the industrial production environment.

[0010] The amplitude variation of each product function component is analyzed to obtain the human voice interference index of each product function component. The human voice interference index is used to represent the probability of the human voice corresponding to the product function component.

[0011] The interference weight of each product function component is determined based on the device characteristic index and human voice interference index of each product function component.

[0012] An interference signal is obtained based on multiple product function components, a monotonic residual signal, and the interference weight of each product function component. The interference signal is used to generate a noise-reduced signal. The noise-reduced signal and the interference signal have the same amplitude but opposite phase. The noise-reduced signal is used to neutralize the noise in the original audio signal.

[0013] In one embodiment, the step of performing local mean decomposition on the original audio signal to obtain multiple product function components and a monotonic residual signal includes:

[0014] The original audio signal is split into multiple signal frames based on a preset frame length;

[0015] Windowing is applied to multiple signal frames to obtain multiple windowed signals;

[0016] Perform Fourier transforms on multiple window signals to obtain multiple signal transformation information;

[0017] Based on the multiple signal transformation information, the signal energy difference between adjacent window signals is analyzed to obtain the decomposition interference index of each window signal, wherein the decomposition interference index is used to represent the probability of the window signal interfering with the local mean decomposition operation.

[0018] Based on the decomposition interference index of each window signal, the original audio signal is decomposed into local mean, resulting in multiple product function components and a monotonic residual signal.

[0019] In one embodiment, the signal transformation information includes multiple signal energies at multiple frequencies in a corresponding window signal;

[0020] The analysis of signal energy differences between adjacent window signals based on the multiple signal transformation information yields the decomposed interference index for each window signal, including:

[0021] Based on the multiple signal transformation information, the energy difference before and energy difference after each frequency in each window signal are obtained, wherein the energy difference before is the signal energy difference between the corresponding frequency in the previous window signal and the corresponding frequency in the corresponding window signal, and the energy difference after is the signal energy difference between the corresponding frequency in the corresponding window signal and the corresponding frequency in the next window signal.

[0022] Calculate the sum of the energy difference before and after each frequency in each window signal to obtain the energy sum value corresponding to each frequency in each window signal;

[0023] Calculate the product of the signal energy and the corresponding sum of energy for each frequency in each window signal to obtain the energy product for each frequency in each window signal;

[0024] Calculate the sum of multiple energy products of multiple frequencies in each window signal to obtain the energy reference value corresponding to each window signal;

[0025] The ratio of the energy reference value to the corresponding energy entropy value for each window signal is calculated to obtain the decomposition interference index for each window signal. The energy entropy value is used to represent the information entropy of multiple signal energies at multiple frequencies in the corresponding window signal.

[0026] In one embodiment, in the multiple decomposition operations corresponding to the local mean decomposition, the extreme value data of each local extreme point in each decomposition operation is the product of the original extreme value of the local extreme point and the correction coefficient corresponding to the local extreme point. The correction coefficient is the ratio of the first interference index and the second interference index of the corresponding local extreme point. The first interference index is the decomposition interference index of the window signal to which the corresponding local extreme point belongs, and the second interference index is the sum of the decomposition interference index of the window signal to which the previous local extreme point belongs and the decomposition interference index of the window signal to which the next local extreme point belongs.

[0027] In one embodiment, the analysis of each product function component based on the original audio signal to obtain the device characteristic index of each product function component includes:

[0028] Perform a fast Fourier transform on each product function component to obtain multiple sinusoidal components corresponding to each product function component;

[0029] Based on the original audio signal, the sinusoidal acoustic index and frequency difference index of each sinusoidal component corresponding to each product function component are determined. The sinusoidal acoustic index is the frequency acoustic index of the frequency component with the smallest frequency difference from the corresponding sinusoidal component among the multiple frequency components included in the original audio signal. The frequency acoustic index is used to indicate the probability that the corresponding frequency component indicates the operating sound of equipment in the corresponding industrial production environment. The frequency difference index is used to indicate the minimum frequency difference between the corresponding sinusoidal component and the multiple frequency components included in the original audio signal.

[0030] Based on the sinusoidal acoustic index, frequency difference index, and amplitude of each sinusoidal component corresponding to each product function component, the component characteristic index of each sinusoidal component corresponding to each product function component is determined.

[0031] Calculate the sum of the component characteristic indices of the multiple sine wave components corresponding to each product function component to obtain the device characteristic index of each product function component.

[0032] In one embodiment, the step of obtaining the frequency acoustic index of each frequency component of the original audio signal includes:

[0033] The original audio signal is split into multiple signal frames based on a preset frame length;

[0034] Windowing is applied to multiple signal frames to obtain multiple windowed signals;

[0035] The amplitude variation of each frequency component in multiple window signals is analyzed to obtain the amplitude fluctuation index corresponding to each frequency component.

[0036] The phase change of each frequency component in multiple window signals is analyzed to obtain the phase change index corresponding to each frequency component.

[0037] The frequency acoustic index of each frequency component is determined based on the amplitude fluctuation index and phase change index corresponding to each frequency component.

[0038] In one embodiment, the analysis of the amplitude variation of each frequency component in multiple window signals to obtain the amplitude fluctuation index corresponding to each frequency component includes:

[0039] Calculate the absolute value of the amplitude difference between two adjacent window signals for each frequency component to obtain multiple window amplitude differences corresponding to each frequency component;

[0040] The variance of the multiple window amplitude differences corresponding to each frequency component is used to obtain the amplitude fluctuation index corresponding to each frequency component.

[0041] In one embodiment, the step of analyzing the phase change of each frequency component in multiple window signals to obtain the phase change index corresponding to each frequency component includes:

[0042] Calculate the absolute value of the phase difference between each frequency component and two adjacent window signals to obtain the phase difference sequence corresponding to each frequency component;

[0043] In the phase difference sequence corresponding to each frequency component, the mean value and multiple peak values ​​of the sequence corresponding to each frequency component are determined, and the difference between the multiple peak values ​​and the mean value of the sequence is calculated to obtain the multiple sequence jump values ​​corresponding to each frequency component.

[0044] Calculate the mean of multiple sequence jump values ​​corresponding to each frequency component to obtain the phase jump index corresponding to each frequency component;

[0045] Curve fitting is performed on the phase difference sequence corresponding to each frequency component to obtain the phase difference fitting curve for each frequency component.

[0046] Calculate the absolute value of the derivative of each sequence point in the phase difference sequence corresponding to each frequency component in the corresponding phase difference fitting curve, and obtain the derivative values ​​of multiple sequence points corresponding to each frequency component.

[0047] The average of the derivative values ​​of multiple sequence points corresponding to each frequency component is determined as the phase drift index corresponding to each frequency component.

[0048] Based on the phase drift index and phase jump index corresponding to each frequency component, the phase change index corresponding to each frequency component is obtained.

[0049] In one embodiment, the analysis of the amplitude variation of each product function component to obtain the human voice interference index for each product function component includes:

[0050] Each product function component is decomposed into multiple signal segments corresponding to each product function component.

[0051] Clustering is performed on multiple signal segments corresponding to each product function component to obtain multiple clusters corresponding to each product function component;

[0052] Using the mean magnitude of the clusters as a comparison index, the multiple clusters corresponding to each product function component are sorted in descending order to obtain the cluster sequence corresponding to each product function component;

[0053] In the cluster sequence corresponding to each product function component, the difference between the mean amplitudes of two adjacent clusters is calculated to obtain multiple cluster amplitude differences corresponding to each product function component, wherein the cluster amplitude difference is the difference between the mean amplitude of the corresponding cluster and its preceding cluster.

[0054] Among the multiple cluster amplitude differences corresponding to each product function component, the cluster corresponding to the largest cluster amplitude difference is determined as the critical cluster corresponding to each product function component, and the multiple clusters located after the critical cluster in the corresponding cluster sequence are determined as the multiple strong signal clusters corresponding to each product function component.

[0055] The amplitude changes among multiple strong signal clusters corresponding to each product function component are analyzed to obtain the human voice interference index for each product function component.

[0056] In one embodiment, the analysis of amplitude changes among multiple strong signal clusters corresponding to each product function component to obtain the human voice interference index for each product function component includes:

[0057] In each strong signal cluster corresponding to each product function component, the degree of cluster difference between each strong signal cluster and its adjacent next strong signal cluster is analyzed to obtain the cluster difference value corresponding to each product function component.

[0058] In each strong signal cluster corresponding to each product function component, the difference between the average cluster amplitude of each strong signal cluster and its adjacent next strong signal cluster is calculated to obtain multiple target amplitude differences corresponding to each product function component.

[0059] Calculate the ratio of the target magnitude difference to the corresponding cluster difference value for each component of the product function to obtain multiple cluster interference indices for each component of the product function.

[0060] Calculate the mean of multiple cluster interference indices corresponding to each product function component to obtain the human voice interference index for each product function component.

[0061] Secondly, another embodiment of the present invention provides a broadband adaptive filtering and inverse acoustic synthesis system for an intelligent noise reduction device, the system comprising:

[0062] The signal decomposition module is used to perform local mean decomposition on the original audio signal to obtain multiple product function components and a monotonic residual signal. The original audio signal originates from an industrial production environment.

[0063] The equipment feature analysis module is used to analyze each product function component based on the original audio signal to obtain the equipment feature index of each product function component. The equipment feature index is used to represent the probability of the product function component corresponding to the equipment operation sound in the industrial production environment.

[0064] The human voice analysis module is used to analyze the amplitude change of each product function component to obtain the human voice interference index of each product function component. The human voice interference index is used to represent the probability of the human voice corresponding to the product function component.

[0065] The interference analysis module is used to determine the interference weight of each product function component based on the device characteristic index and human voice interference index of each product function component.

[0066] An interference determination module is used to obtain an interference signal based on multiple product function components, a monotonic residual signal, and the interference weight of each product function component. The interference signal is used to generate a noise reduction signal. The noise reduction signal and the interference signal have the same amplitude but opposite phase. The noise reduction signal is used to neutralize the noise in the original audio signal.

[0067] Thirdly, in another embodiment of the present invention, an electronic device is provided, including a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the method described in the first aspect.

[0068] Fourthly, in another embodiment of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the method described in the first aspect.

[0069] The present invention has the following beneficial effects:

[0070] This invention performs local mean decomposition on the original audio signal to adapt to the non-stationary, wideband, and nonlinear characteristics of original audio signals from industrial production environments. It accurately decomposes the original audio signal into multiple product function components and a monotonic residual signal. Then, it obtains the equipment characteristic index and human voice interference index for each product function component to determine the probability that each product function component corresponds to equipment operation sound and human voice, i.e., the probability that each product function component belongs to the key sound that needs to be retained and the probability that each product function component belongs to the noise that needs to be filtered out. This allows for a comprehensive and accurate assessment of the interference weight of each product function component, and based on this, a more accurate interference signal can be determined. In other words, it accurately identifies environmental noise and human voice interference that overlap with equipment operation sounds in industrial production environments, thereby improving the noise reduction effect on industrial audio. Attached Figure Description

[0071] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0072] Figure 1 This is a schematic flowchart illustrating a broadband adaptive filtering and reverse acoustic wave synthesis method for an intelligent noise reduction device according to an embodiment of the present invention.

[0073] Figure 2This is a schematic diagram of the structure of a broadband adaptive filtering and inverse acoustic synthesis system for an intelligent noise reduction device provided in an embodiment of the present invention.

[0074] Figure 3 This is a schematic diagram of the structure of an electronic device provided in one embodiment of the present invention. Detailed Implementation

[0075] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a broadband adaptive filtering and reverse acoustic wave synthesis method for an intelligent noise reduction device proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0076] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0077] The following description, in conjunction with the accompanying drawings, details the specific scheme of the broadband adaptive filtering and reverse acoustic wave synthesis method for an intelligent noise reduction device provided by the present invention.

[0078] This invention proposes a broadband adaptive filtering and reverse acoustic wave synthesis method for intelligent noise reduction devices. Please refer to [link to relevant documentation]. Figure 1 The diagram illustrates a schematic flowchart of a broadband adaptive filtering and reverse acoustic wave synthesis method for an intelligent noise reduction device according to an embodiment of the present invention. The method includes the following steps:

[0079] Step S1: Perform local mean decomposition on the original audio signal to obtain multiple product function components and a monotonic residual signal.

[0080] The original audio signal originated from an industrial production environment.

[0081] In industrial production environments, the sources of raw audio signals are complex, including not only continuous equipment operating sounds (such as motor sounds, fan sounds, and mechanical vibration sounds) and environmental noise, but also sudden equipment start-up sounds, impact sounds, operation alarm sounds, and human voices. The frequency distribution of these sounds typically spans from 20 Hz to 16 kHz, and their spectral characteristics change in real time.

[0082] In applications, to effectively acquire raw audio signals containing device operating sounds, ambient noise, and sudden noises, the following setup can be implemented:

[0083] A multi-channel microphone array (generally ≥4 channels) is deployed near key fixed components of the equipment (such as motor bearings, gearboxes, pumps, etc.). The array structure of the multi-channel microphone array can be adjusted to linear, circular, or area array according to the spatial layout of the production site. Each microphone collects the on-site sound signal at a sampling rate of 44.1kHz to 192kHz, with a quantization accuracy of ≥16bit and a dynamic range of ≥96dB. All microphones are sampled synchronously with timestamps to ensure the integrity of the multi-channel audio phase information.

[0084] The original audio signal can be acquired using the multi-channel microphone array described above.

[0085] In the above steps, the environmental noise and equipment operating sound are separated into different components by decomposing the original audio signal. This allows for adaptive suppression of noise in different components based on their different characteristics, while preserving the characteristics of the equipment operating sound as much as possible.

[0086] The reason for choosing local mean decomposition is that the original audio signal in the industrial production environment has non-stationary, wideband, and nonlinear characteristics, and is mixed with complex components such as background noise (fan and pipe noise), human voices, sudden impact noise, and vibration noise of the equipment's natural operating frequency. Local mean decomposition can utilize the local characteristics of the original audio signal itself to decompose the audio, which is better suited to the non-stationary, wideband, and nonlinear characteristics of the original audio signal. This ensures the accuracy of the multiple product function components and the monotonic residual signal obtained from the decomposition, thereby ensuring the accuracy of the noise signal determined subsequently.

[0087] Furthermore, the step of performing local mean decomposition on the original audio signal to obtain multiple product function components and a monotonic residual signal includes:

[0088] The original audio signal is split into multiple signal frames based on a preset frame length;

[0089] Windowing is applied to multiple signal frames to obtain multiple windowed signals;

[0090] Perform Fourier transforms on multiple window signals to obtain multiple signal transformation information;

[0091] Based on the multiple signal transformation information, the signal energy difference between adjacent window signals is analyzed to obtain the decomposition interference index of each window signal, wherein the decomposition interference index is used to represent the probability of the window signal interfering with the local mean decomposition operation.

[0092] Based on the decomposition interference index of each window signal, the original audio signal is decomposed into local mean, resulting in multiple product function components and a monotonic residual signal.

[0093] For example, the frame length can be 50 milliseconds (ms), and the window function used in the windowing process can be a Hamming window function.

[0094] Each window signal, after undergoing a Fourier transform, will generate a signal transformation information that includes several sine waves with different amplitudes, frequencies, and phases corresponding to that window signal.

[0095] Each window signal contains multiple sine waves, and these sine waves have different frequencies. Therefore, it can also be understood that each window signal contains multiple frequencies, and the amplitude and phase of each frequency are specifically the amplitude and phase of the corresponding sine wave.

[0096] When there are sudden impact sounds in the industrial field (such as hammering, valve bursting, accidental operation), these strong abrupt changes will cause the audio extreme points to increase sharply in a short period of time, and the energy is high. At this time, if local mean decomposition is performed directly, there will be a high risk of fitting distortion, resulting in PF decomposition errors, pseudo-modes or mode mixing, which will affect the determination of subsequent noise signals.

[0097] Based on this, in the above process, the signal energy difference between adjacent window signals is analyzed to obtain the decomposition interference index of each window signal, that is, to quantify the probability of sudden impact sound in each signal segment of the original audio signal, so as to guide the subsequent local mean decomposition operation, suppress the adverse effects of strong abrupt signal caused by sudden impact sound on the local mean decomposition operation, so as to make the multiple product function components and a monotonic residual signal obtained after decomposition more accurate and reliable.

[0098] Furthermore, the signal transformation information includes multiple signal energies at multiple frequencies in the corresponding window signal;

[0099] The analysis of signal energy differences between adjacent window signals based on the multiple signal transformation information yields the decomposed interference index for each window signal, including:

[0100] Based on the multiple signal transformation information, the energy difference before and energy difference after each frequency in each window signal are obtained, wherein the energy difference before is the signal energy difference between the corresponding frequency in the previous window signal and the corresponding frequency in the corresponding window signal, and the energy difference after is the signal energy difference between the corresponding frequency in the corresponding window signal and the corresponding frequency in the next window signal.

[0101] Calculate the sum of the energy difference before and after each frequency in each window signal to obtain the energy sum value corresponding to each frequency in each window signal;

[0102] Calculate the product of the signal energy and the corresponding sum of energy for each frequency in each window signal to obtain the energy product for each frequency in each window signal;

[0103] Calculate the sum of multiple energy products of multiple frequencies in each window signal to obtain the energy reference value corresponding to each window signal;

[0104] The ratio of the energy reference value to the corresponding energy entropy value for each window signal is calculated to obtain the decomposition interference index for each window signal. The energy entropy value is used to represent the information entropy of multiple signal energies at multiple frequencies in the corresponding window signal.

[0105] Specifically, the signal energy difference between the corresponding frequency in the previous window signal and the corresponding frequency in the corresponding window signal is the difference between the signal energy of the corresponding frequency in the previous window signal and the signal energy of the corresponding frequency in the corresponding window signal. The signal energy can be understood as the square of the amplitude of the corresponding frequency.

[0106] Similarly, the signal energy difference between the corresponding frequency in the corresponding window signal and the corresponding frequency in the subsequent window signal is specifically the difference between the signal energy of the corresponding frequency in the corresponding window signal and the signal energy of the corresponding frequency in the subsequent window signal.

[0107] Among them, the energy reference value corresponding to the window signal is used to represent the degree of signal energy fluctuation in the corresponding window signal (usually manifested as rapid signal energy decay). The larger the energy reference value, the more intense the signal energy fluctuation in the corresponding window signal, which means that the probability of the corresponding window signal containing sudden impact sound (represented by decomposing the interference index) is higher.

[0108] Similarly, the larger the energy entropy value corresponding to the window signal, the greater the information entropy of multiple signal energies at multiple frequencies in the window signal. In other words, the more different frequency audio signals are contained in the window signal, the higher the probability that the window signal contains other noises besides the sound of the device running. Therefore, the probability that the window signal contains sudden impact sounds is lower.

[0109] In applications, to avoid the influence of extreme values, the energy difference before and after each frequency in each window signal can be normalized (e.g., using the sigmoid function) to normalize the energy difference before and after each frequency in each window signal to the (0,1) interval.

[0110] For example, the process of obtaining the information entropy of multiple signal energies at multiple frequencies in a window signal can be as follows:

[0111] Calculate the sum of the energies of multiple signals at multiple frequencies in the window signal to obtain the total energy value corresponding to the window signal;

[0112] Calculate the ratio of the signal energy of each frequency in the windowed signal to its corresponding total energy value to obtain the signal energy proportion of each frequency in the windowed signal. The signal energy proportion of each frequency in the windowed signal can be expressed as... , (j=1,…,num), where num is the total number of frequencies in the window signal;

[0113] Correspondingly, the information entropy of multiple signal energies at multiple frequencies in the window signal. It can be represented as:

[0114]

[0115] In one example, the decomposition interference index of a window signal can be expressed as:

[0116]

[0117] in, This represents the decomposition interference index of the windowed signal, where n represents the total number of frequencies in the windowed signal. This represents the normalized value of the signal energy of the sine wave at the i-th frequency in the window signal (the normalization process can be completed using the Z-Score normalization algorithm). and This represents the energy difference before and after the sine wave of the i-th frequency in the window signal. This represents the energy entropy value corresponding to the window signal.

[0118] Furthermore, in the multiple decomposition operations corresponding to the local mean decomposition, the extreme value data of each local extreme point in each decomposition operation is the product of the original extreme value of the local extreme point and the correction coefficient corresponding to the local extreme point. The correction coefficient is the ratio of the first interference index and the second interference index of the corresponding local extreme point. The first interference index is the decomposition interference index of the window signal to which the corresponding local extreme point belongs, and the second interference index is the sum of the decomposition interference index of the window signal to which the previous local extreme point belongs and the decomposition interference index of the window signal to which the next local extreme point belongs.

[0119] The original extreme value of a local extreme point can be understood as the original amplitude of the local extreme point.

[0120] Specifically, the multiple decomposition operations corresponding to local mean decomposition are as follows:

[0121] Step 1: Extract multiple local extrema (including local maxima and local minima) from the original audio signal.

[0122] Step 2: Among the above multiple local extrema, calculate the product of the magnitude of each local extrema and its corresponding correction coefficient to obtain the correction magnitude of each local extrema.

[0123] Step 3: Among the above multiple local extreme points, the correction magnitudes of adjacent local extreme points are averaged to obtain the local mean.

[0124] Step 4: Then, among the above multiple local extrema, calculate half the absolute value of the magnitude difference between adjacent local extrema to obtain the envelope estimate.

[0125] Step 5: Smooth the local mean and envelope estimates using the moving average method to obtain the smoothed local mean function. and local envelope estimation function ;

[0126] Step 6: Subtract the local mean function from the original audio signal. To obtain a zero-mean signal ;

[0127] Step 7, for Demodulation is performed, specifically through... Divide by To obtain the envelope normalized signal ,

[0128] Step 8: Repeat steps 1 to 7 above until the local envelope estimation function is obtained. The envelope-normalized signal obtained at this time is a pure frequency-modulated signal, which can be defined as follows: It should be noted that the multiple local extrema used in each decomposition operation need to be extracted from the zero-mean signal obtained in the previous decomposition operation. The multiple local extrema used in the first decomposition operation need to be obtained from the original audio signal. In addition, the audio signal used to subtract the local mean function in each decomposition operation is specifically the zero-mean signal obtained in the previous decomposition operation.

[0129] Step 9: Multiply all the local envelope estimation functions obtained from the above multiple decomposition operations to obtain the envelope signal. Then the first component of the product function (i.e., the PF component) can be expressed as:

[0130] Step 10: Subtract from the original audio signal The remaining signal is then repeated through steps one through nine (it should be noted that during the repeated execution of the above steps, the original audio signal in step one will be changed to the most recently obtained remaining signal, and similarly, the original audio signal in step six will be changed to the most recently obtained remaining signal minus the zero mean function obtained by the multiple local mean functions), until the remaining signal is a monotonic function (i.e., a monotonic remaining signal). At this point, the original audio signal is decomposed into several PF components and a monotonic remaining signal, thus completing the decomposition.

[0131] In the above process, the correction coefficient for each local extremum point is calculated based on the decomposition interference index of each window signal. This allows for adaptive correction of the amplitude of each local extremum point when participating in the mean and envelope estimation calculations, based on the probability that the window signal to which each local extremum point belongs contains sudden impact sound. This suppresses the influence of extreme amplitudes of local extremum points caused by sudden impact sound, avoids the risk of fitting distortion in the local mean estimation operation, and makes the obtained multiple product function components and a monotonic residual signal more accurate.

[0132] Step S2: Analyze each product function component based on the original audio signal to obtain the device characteristic index of each product function component.

[0133] The equipment characteristic index is used to represent the probability of the product function component corresponding to the operating sound of equipment in the industrial production environment.

[0134] Furthermore, the analysis of each product function component based on the original audio signal to obtain the device characteristic index of each product function component includes:

[0135] Perform a fast Fourier transform on each product function component to obtain multiple sinusoidal components corresponding to each product function component;

[0136] Based on the original audio signal, the sinusoidal acoustic index and frequency difference index of each sinusoidal component corresponding to each product function component are determined. The sinusoidal acoustic index is the frequency acoustic index of the frequency component with the smallest frequency difference from the corresponding sinusoidal component among the multiple frequency components included in the original audio signal. The frequency acoustic index is used to indicate the probability that the corresponding frequency component indicates the operating sound of equipment in the corresponding industrial production environment. The frequency difference index is used to indicate the minimum frequency difference between the corresponding sinusoidal component and the multiple frequency components included in the original audio signal.

[0137] Based on the sinusoidal acoustic index, frequency difference index, and amplitude of each sinusoidal component corresponding to each product function component, the component characteristic index of each sinusoidal component corresponding to each product function component is determined.

[0138] Calculate the sum of the component characteristic indices of the multiple sine wave components corresponding to each product function component to obtain the device characteristic index of each product function component.

[0139] In the above process, based on the sine wave acoustic index, frequency difference index, and amplitude of each sine wave component corresponding to each product function component, the component characteristic index of each sine wave component corresponding to each product function component is calculated. That is, by using the amplitude of each sine wave, the probability that the most similar frequency of the sine wave in the original audio signal indicates the sound of the device, and the degree of difference (or similarity) of the most similar frequency of the sine wave in the original audio signal, the probability that each sine wave component corresponding to each product function component indicates the sound of the device is comprehensively evaluated. Then, by summing up the multiple probabilities of multiple sine wave components of each product function component, a more accurate device characteristic index can be obtained.

[0140] For example, a device characteristic index of a product function component It can be represented as:

[0141]

[0142] Where M represents the total number of sinusoidal components included in the product function. This represents the amplitude of the m-th sine wave component in the product function (the amplitude can be normalized to avoid the influence of extreme values). This represents the sinusoidal acoustic index of the m-th sinusoidal component in the product function. It represents the frequency difference index of the m-th sine wave component in the product function.

[0143] Furthermore, the steps for obtaining the frequency acoustic index of each frequency component of the original audio signal include:

[0144] The original audio signal is split into multiple signal frames based on a preset frame length;

[0145] Windowing is applied to multiple signal frames to obtain multiple windowed signals;

[0146] The amplitude variation of each frequency component in multiple window signals is analyzed to obtain the amplitude fluctuation index corresponding to each frequency component.

[0147] The phase change of each frequency component in multiple window signals is analyzed to obtain the phase change index corresponding to each frequency component.

[0148] The frequency acoustic index of each frequency component is determined based on the amplitude fluctuation index and phase change index corresponding to each frequency component.

[0149] In this process, the amplitude and phase changes of each frequency component in the original audio signal are analyzed in multiple window signals to evaluate the amplitude fluctuation and phase change of each frequency component. Based on this, the frequency acoustic index of each frequency component in the original audio signal is determined, which can more accurately assess the probability of the device operating sound corresponding to each frequency component.

[0150] The larger the amplitude fluctuation index, the more drastic the amplitude change of the corresponding frequency component among multiple window signals, and the less it matches the stable sound characteristics generated during equipment operation. Therefore, the probability that the corresponding frequency component indicates the operating sound of the equipment (specifically, the normal operating condition) is lower.

[0151] The larger the phase change index, the more significant the phase change of the corresponding frequency among multiple window signals, and the better it matches the phase trajectory anomalies (such as phase jumps and phase drifts) that occur during abnormal equipment operation. In this case, the probability that the corresponding frequency component indicates the sound of equipment operation (specifically, abnormal operation conditions, such as bearing wear, gear cracks, etc.) is higher.

[0152] For example, the frequency acoustic index of the frequency component It can be represented as:

[0153]

[0154] in, The phase change index used to represent this frequency component. The amplitude fluctuation index used to represent this frequency component.

[0155] Specifically, the analysis of the amplitude variation of each frequency component in multiple window signals to obtain the amplitude fluctuation index corresponding to each frequency component includes:

[0156] Calculate the absolute value of the amplitude difference between two adjacent window signals for each frequency component to obtain multiple window amplitude differences corresponding to each frequency component;

[0157] The variance of the multiple window amplitude differences corresponding to each frequency component is used to obtain the amplitude fluctuation index corresponding to each frequency component.

[0158] In this process, choosing to quantify the amplitude fluctuation of each frequency component in multiple window signals by using the variance of the amplitude difference between adjacent window signals for each frequency component, rather than the variance of the amplitude of each frequency component in multiple window signals, can make the determined amplitude fluctuation index more accurate.

[0159] Specifically, the analysis of the phase change of each frequency component in multiple window signals to obtain the phase change index corresponding to each frequency component includes:

[0160] Calculate the absolute value of the phase difference between each frequency component and two adjacent window signals to obtain the phase difference sequence corresponding to each frequency component;

[0161] In the phase difference sequence corresponding to each frequency component, the mean value and multiple peak values ​​of the sequence corresponding to each frequency component are determined, and the difference between the multiple peak values ​​and the mean value of the sequence is calculated to obtain the multiple sequence jump values ​​corresponding to each frequency component.

[0162] Calculate the mean of multiple sequence jump values ​​corresponding to each frequency component to obtain the phase jump index corresponding to each frequency component;

[0163] Curve fitting is performed on the phase difference sequence corresponding to each frequency component to obtain the phase difference fitting curve for each frequency component.

[0164] Calculate the absolute value of the derivative of each sequence point in the phase difference sequence corresponding to each frequency component in the corresponding phase difference fitting curve, and obtain the derivative values ​​of multiple sequence points corresponding to each frequency component.

[0165] The average of the derivative values ​​of multiple sequence points corresponding to each frequency component is determined as the phase drift index corresponding to each frequency component.

[0166] Based on the phase drift index and phase jump index corresponding to each frequency component, the phase change index corresponding to each frequency component is obtained.

[0167] Wherein, the sequence mean is the average value of multiple sequence elements included in the corresponding phase difference sequence, the sequence peak value is the value of the peak element among the multiple sequence elements included in the corresponding phase difference sequence, and the peak element is the sequence element among the multiple sequence elements included in the corresponding phase difference sequence that is greater than the previous sequence element and greater than the next sequence element.

[0168] The phase jump index is used to indicate the degree to which the phase peak value of the corresponding frequency component deviates from its phase mean value. The larger the phase jump index, the more significant the phase jump characteristic of the corresponding frequency component, and the better it matches the phase trajectory change that occurs during abnormal operation of the equipment.

[0169] The phase drift index is used to represent the degree of phase shift in the corresponding frequency component. The larger the phase drift index, the more significant the phase drift characteristic of the corresponding frequency component, and the better it matches the phase trajectory change that occurs during abnormal operation of the equipment.

[0170] In applications, the phase value of the frequency component in the window signal It can be represented as:

[0171]

[0172] in, This represents the phase angle value of the frequency component in the window signal, and cos(.) represents the cosine function.

[0173] In one example, the product of the phase drift index and the phase jump index corresponding to each frequency component can be calculated to obtain the phase change index corresponding to each frequency component. In this case, the above curve fitting process can be completed based on the least squares method.

[0174] Step S3: Analyze the amplitude change of each product function component to obtain the human voice interference index of each product function component.

[0175] The human voice interference index is used to represent the probability of a human voice corresponding to a component of the product function.

[0176] Furthermore, the analysis of the amplitude variation of each product function component yields the human voice interference index for each product function component, including:

[0177] Each product function component is decomposed into multiple signal segments corresponding to each product function component.

[0178] Clustering is performed on multiple signal segments corresponding to each product function component to obtain multiple clusters corresponding to each product function component;

[0179] Using the mean magnitude of the clusters as a comparison index, the multiple clusters corresponding to each product function component are sorted in descending order to obtain the cluster sequence corresponding to each product function component;

[0180] In the cluster sequence corresponding to each product function component, the difference between the mean amplitudes of two adjacent clusters is calculated to obtain multiple cluster amplitude differences corresponding to each product function component, wherein the cluster amplitude difference is the difference between the mean amplitude of the corresponding cluster and its preceding cluster.

[0181] Among the multiple cluster amplitude differences corresponding to each product function component, the cluster corresponding to the largest cluster amplitude difference is determined as the critical cluster corresponding to each product function component, and the multiple clusters located after the critical cluster in the corresponding cluster sequence are determined as the multiple strong signal clusters corresponding to each product function component.

[0182] The amplitude changes among multiple strong signal clusters corresponding to each product function component are analyzed to obtain the human voice interference index for each product function component.

[0183] For example, each product function component can be split into multiple signal segments based on a preset segment length to obtain multiple signal segments corresponding to each product function component, wherein the segment length can be 50 milliseconds.

[0184] The above clustering process can be completed using the K-Means algorithm. The K value of the K-Means algorithm can be obtained by the elbow method. The average value of multiple signal amplitudes in the corresponding signal segment and the start time of the signal segment (that is, the time indicated by the first signal point of the signal segment) can be used as the clustering index. That is to say, the average amplitude values ​​(that is, the average value of multiple signal amplitudes in the signal segment) of different signal segments in the same cluster are close, and the start times are also close.

[0185] In the above process, the mean amplitude of the clusters is selected as the comparison index. The multiple clusters corresponding to each product function component are sorted in descending order to obtain the cluster sequence corresponding to each product function component. The cluster with the largest difference in mean amplitude among the clusters is selected as the critical cluster. In order to determine several strong signal clusters corresponding to strong signals from multiple clusters, the interference of corresponding low signal energy noise can be avoided as much as possible, so that the determined human voice interference index is more accurate.

[0186] Furthermore, the analysis of the amplitude changes among multiple strong signal clusters corresponding to each product function component to obtain the human voice interference index for each product function component includes:

[0187] In each strong signal cluster corresponding to each product function component, the degree of cluster difference between each strong signal cluster and its adjacent next strong signal cluster is analyzed to obtain the cluster difference value corresponding to each product function component.

[0188] In each strong signal cluster corresponding to each product function component, the difference between the average cluster amplitude of each strong signal cluster and its adjacent next strong signal cluster is calculated to obtain multiple target amplitude differences corresponding to each product function component.

[0189] Calculate the ratio of the target magnitude difference to the corresponding cluster difference value for each component of the product function to obtain multiple cluster interference indices for each component of the product function.

[0190] Calculate the mean of multiple cluster interference indices corresponding to each product function component to obtain the human voice interference index for each product function component.

[0191] Among them, the average amplitude of the strong signal cluster is the average of the average amplitudes of several signal segments included in the strong signal cluster.

[0192] In the above process, the degree of cluster difference and the difference in the mean cluster amplitude between each strong signal cluster and its adjacent next strong signal cluster are analyzed. This allows for an accurate assessment of the probability of human voice interference corresponding to each strong signal cluster from both the differences between the clusters as a whole and the differences in the amplitude dimension, making the determined human voice interference index more accurate.

[0193] The larger the target amplitude difference, the greater the increase in signal amplitude (or signal energy) between two adjacent strong signal clusters. This also matches the situation in industrial production environments where human voices gradually increase (due to the strong background noise, human voice communication needs to be repeated multiple times and gradually increase). Therefore, the probability that the corresponding product function component indicates human voices is also greater.

[0194] The larger the cluster difference value, the lower the similarity between two adjacent strong signal clusters, the lower the matching degree with the repetition of human voices in the industrial production environment, and the lower the probability that the corresponding product function component indicates human voices.

[0195] For example, the degree of cluster difference between each strong signal cluster and its adjacent next strong signal cluster can be quantified by calculating the dynamic time warping distance between each strong signal cluster and its next adjacent strong signal cluster, thereby obtaining multiple cluster difference values ​​corresponding to each product function component. The target amplitude difference is specifically the difference between the mean cluster amplitude of the next adjacent cluster and the mean cluster amplitude of the corresponding cluster.

[0196] In one example, the human voice interference index of the product function component. It can be represented as:

[0197]

[0198] Where U represents the total number of strong signal clusters included in the product function components. This represents the target amplitude difference of the r-th strong signal cluster included in the product function component (i.e., the difference between the average cluster amplitude of the (r+1)-th strong signal cluster included in the product function component and the average cluster amplitude of the r-th strong signal cluster). This represents the cluster difference value of the r-th strong signal cluster included in the product function component (i.e., the dynamic time warping distance between the r-th strong signal cluster and the (r+1)-th strong signal cluster included in the product function component).

[0199] Step S4: Determine the interference weight of each product function component based on the device characteristic index and human voice interference index of each product function component.

[0200] Specifically, the initial weight of each product function component can be obtained by calculating the ratio of the human voice interference index to the device characteristic index for each product function component.

[0201] By normalizing the initial weights of each product function component, the interference weights of each product function component can be obtained.

[0202] Step S5: Obtain the interference signal based on multiple product function components, the monotonic residual signal, and the interference weight of each product function component.

[0203] The interference signal is used to generate a noise-reduced signal. The noise-reduced signal and the interference signal have the same amplitude but opposite phase. The noise-reduced signal is used to neutralize the noise in the original audio signal.

[0204] In one example, multiple product function components can be weighted based on the interference weight of each product function component to obtain a weighted average signal; then the enhanced average signal and the monotonic residual signal are added together to obtain the interference signal.

[0205] It should be noted that since the monotonic residual signal does not possess the non-stationary, wideband, and nonlinear characteristics of equipment operating sound, it can be directly regarded as environmental noise and used as the basic component signal of the interference signal.

[0206] In applications, the denoised signal and the original audio signal can be added together to obtain the target audio signal after adaptive noise reduction processing. This target audio signal can be used to analyze the operating status and fault risks of various devices in an industrial production environment. This invention does not limit the application of the target audio signal.

[0207] In summary, this invention performs local mean decomposition on the original audio signal to adapt to the non-stationary, wideband, and nonlinear characteristics of the original audio signal originating from an industrial production environment, obtaining multiple product function components and a monotonic residual signal. Then, it acquires the device characteristic index and human voice interference index for each product function component to determine the probability that each product function component corresponds to equipment operation sound and the probability that it corresponds to human voice. This means determining the probability that each product function component belongs to the key sound that needs to be retained and the probability that it belongs to the noise that needs to be filtered out. This allows for a more accurate assessment of the interference weight of each product function component from multiple perspectives, and based on this, a more accurate interference signal can be determined. In other words, it accurately identifies environmental noise and human voice interference that overlap with equipment operation sounds in an industrial production environment, thereby improving the noise reduction effect on industrial audio.

[0208] This invention proposes a broadband adaptive filtering and inverse acoustic synthesis system for intelligent noise reduction devices. Please refer to [link / reference]. Figure 2 The diagram illustrates a schematic of a broadband adaptive filtering and inverse acoustic synthesis system 200 for an intelligent noise reduction device according to an embodiment of the present invention. The system includes:

[0209] The signal decomposition module 201 is used to perform local mean decomposition on the original audio signal to obtain multiple product function components and a monotonic residual signal. The original audio signal originates from an industrial production environment.

[0210] The equipment feature analysis module 202 is used to analyze each product function component based on the original audio signal to obtain the equipment feature index of each product function component. The equipment feature index is used to represent the probability of the product function component corresponding to the equipment operation sound in the industrial production environment.

[0211] The human voice analysis module 203 is used to analyze the amplitude change of each product function component to obtain the human voice interference index of each product function component. The human voice interference index is used to represent the probability of the human voice corresponding to the product function component.

[0212] Interference analysis module 204 is used to determine the interference weight of each product function component based on the device characteristic index and human voice interference index of each product function component.

[0213] The interference determination module 205 is used to obtain an interference signal based on multiple product function components, a monotonic residual signal, and the interference weight of each product function component. The interference signal is used to generate a noise reduction signal. The noise reduction signal and the interference signal have the same amplitude but opposite phase. The noise reduction signal is used to neutralize the noise in the original audio signal.

[0214] It should be noted that the system provided in the above embodiments is only an example of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the broadband adaptive filtering and inverse acoustic synthesis system of an intelligent noise reduction device provided in the above embodiments and the broadband adaptive filtering and inverse acoustic synthesis method embodiment of an intelligent noise reduction device belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0215] This invention also provides an electronic device. Please refer to [link to relevant documentation]. Figure 3 The electronic device may include a processor 301, a memory 302, and a program 3021 stored in the memory 302 and capable of running on the processor 301.

[0216] When program 3021 is executed by processor 301, it can achieve the following: Figure 1 Any steps in the corresponding method embodiments and the achievement of the same beneficial effects will not be repeated here.

[0217] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by hardware related to program instructions, and the program can be stored in a readable medium.

[0218] This invention also provides a readable storage medium storing a computer program, which, when executed by a processor, can perform the above-described functions. Figure 1 Any step in the corresponding method embodiment can achieve the same technical effect, and will not be repeated here to avoid repetition.

[0219] The computer-readable storage medium of this invention can be any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. For example, a computer-readable storage medium can be an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0220] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0221] The program code contained on the storage medium can be transmitted using any suitable medium, including but not limited to wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0222] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or terminal. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0223] This invention also provides a computer program product that, when run on a computer, causes the computer to perform the aforementioned steps to implement the broadband adaptive filtering and reverse acoustic wave synthesis method for an intelligent noise reduction device provided in the above embodiments.

[0224] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0225] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A broadband adaptive filtering and reverse acoustic wave synthesis method for an intelligent noise reduction device, characterized in that, The method includes: The original audio signal is subjected to local mean decomposition to obtain multiple product function components and a monotonic residual signal. The original audio signal is derived from an industrial production environment. Based on the analysis of each product function component of the original audio signal, the device characteristic index of each product function component is obtained. The device characteristic index is used to represent the probability of the product function component corresponding to the sound of equipment operation in the industrial production environment. The amplitude variation of each product function component is analyzed to obtain the human voice interference index of each product function component. The human voice interference index is used to represent the probability of the human voice corresponding to the product function component. The interference weight of each product function component is determined based on the device characteristic index and human voice interference index of each product function component. An interference signal is obtained based on multiple product function components, a monotonic residual signal, and the interference weight of each product function component. The interference signal is used to generate a noise-reduced signal. The noise-reduced signal and the interference signal have the same amplitude but opposite phase. The noise-reduced signal is used to neutralize the noise in the original audio signal.

2. The broadband adaptive filtering and reverse acoustic wave synthesis method for intelligent noise reduction equipment according to claim 1, characterized in that, The process of performing local mean decomposition on the original audio signal to obtain multiple product function components and a monotonic residual signal includes: The original audio signal is split into multiple signal frames based on a preset frame length; Windowing is applied to multiple signal frames to obtain multiple windowed signals; Perform Fourier transforms on multiple window signals to obtain multiple signal transformation information; Based on the multiple signal transformation information, the signal energy difference between adjacent window signals is analyzed to obtain the decomposition interference index of each window signal, wherein the decomposition interference index is used to represent the probability of the window signal interfering with the local mean decomposition operation. Based on the decomposition interference index of each window signal, the original audio signal is decomposed into local mean, resulting in multiple product function components and a monotonic residual signal.

3. The broadband adaptive filtering and reverse acoustic wave synthesis method for intelligent noise reduction equipment according to claim 2, characterized in that, The signal transformation information includes multiple signal energies at multiple frequencies in the corresponding window signal; The analysis of signal energy differences between adjacent window signals based on the multiple signal transformation information yields the decomposed interference index for each window signal, including: Based on the multiple signal transformation information, the energy difference before and energy difference after each frequency in each window signal are obtained, wherein the energy difference before is the signal energy difference between the corresponding frequency in the previous window signal and the corresponding frequency in the corresponding window signal, and the energy difference after is the signal energy difference between the corresponding frequency in the corresponding window signal and the corresponding frequency in the next window signal. Calculate the sum of the energy difference before and after each frequency in each window signal to obtain the energy sum value corresponding to each frequency in each window signal; Calculate the product of the signal energy and the corresponding sum of energy for each frequency in each window signal to obtain the energy product for each frequency in each window signal; Calculate the sum of multiple energy products of multiple frequencies in each window signal to obtain the energy reference value corresponding to each window signal; The ratio of the energy reference value to the corresponding energy entropy value for each window signal is calculated to obtain the decomposition interference index for each window signal. The energy entropy value is used to represent the information entropy of multiple signal energies at multiple frequencies in the corresponding window signal.

4. The broadband adaptive filtering and reverse acoustic wave synthesis method for intelligent noise reduction equipment according to claim 2, characterized in that, In the multiple decomposition operations corresponding to the local mean decomposition, the extreme value data of each local extreme point in each decomposition operation is the product of the original extreme value of the local extreme point and the correction coefficient corresponding to the local extreme point. The correction coefficient is the ratio of the first interference index and the second interference index of the corresponding local extreme point. The first interference index is the decomposition interference index of the window signal to which the corresponding local extreme point belongs, and the second interference index is the sum of the decomposition interference index of the window signal to which the previous local extreme point belongs and the decomposition interference index of the window signal to which the next local extreme point belongs.

5. The broadband adaptive filtering and reverse acoustic wave synthesis method for intelligent noise reduction equipment according to claim 1, characterized in that, The analysis of each product function component based on the original audio signal to obtain the device characteristic index of each product function component includes: Perform a fast Fourier transform on each product function component to obtain multiple sinusoidal components corresponding to each product function component; Based on the original audio signal, the sinusoidal acoustic index and frequency difference index of each sinusoidal component corresponding to each product function component are determined. The sinusoidal acoustic index is the frequency acoustic index of the frequency component with the smallest frequency difference from the corresponding sinusoidal component among the multiple frequency components included in the original audio signal. The frequency acoustic index is used to indicate the probability that the corresponding frequency component indicates the operating sound of equipment in the corresponding industrial production environment. The frequency difference index is used to indicate the minimum frequency difference between the corresponding sinusoidal component and the multiple frequency components included in the original audio signal. Based on the sinusoidal acoustic index, frequency difference index, and amplitude of each sinusoidal component corresponding to each product function component, the component characteristic index of each sinusoidal component corresponding to each product function component is determined. Calculate the sum of the component characteristic indices of the multiple sine wave components corresponding to each product function component to obtain the device characteristic index of each product function component.

6. The broadband adaptive filtering and reverse acoustic wave synthesis method for intelligent noise reduction equipment according to claim 5, characterized in that, The steps for obtaining the frequency acoustic index of each frequency component in the original audio signal include: The original audio signal is split into multiple signal frames based on a preset frame length; Windowing is applied to multiple signal frames to obtain multiple windowed signals; The amplitude variation of each frequency component in multiple window signals is analyzed to obtain the amplitude fluctuation index corresponding to each frequency component. The phase change of each frequency component in multiple window signals is analyzed to obtain the phase change index corresponding to each frequency component. The frequency acoustic index of each frequency component is determined based on the amplitude fluctuation index and phase change index corresponding to each frequency component.

7. The broadband adaptive filtering and reverse acoustic wave synthesis method for intelligent noise reduction equipment according to claim 6, characterized in that, The analysis of the amplitude variation of each frequency component in multiple window signals to obtain the amplitude fluctuation index corresponding to each frequency component includes: Calculate the absolute value of the amplitude difference between two adjacent window signals for each frequency component to obtain multiple window amplitude differences corresponding to each frequency component; The variance of the multiple window amplitude differences corresponding to each frequency component is used to obtain the amplitude fluctuation index corresponding to each frequency component.

8. The broadband adaptive filtering and reverse acoustic wave synthesis method for intelligent noise reduction equipment according to claim 6, characterized in that, The analysis of the phase change of each frequency component in multiple window signals to obtain the phase change index corresponding to each frequency component includes: Calculate the absolute value of the phase difference between each frequency component and two adjacent window signals to obtain the phase difference sequence corresponding to each frequency component; In the phase difference sequence corresponding to each frequency component, the mean value and multiple peak values ​​of the sequence corresponding to each frequency component are determined, and the difference between the multiple peak values ​​and the mean value of the sequence is calculated to obtain the multiple sequence jump values ​​corresponding to each frequency component. Calculate the mean of multiple sequence jump values ​​corresponding to each frequency component to obtain the phase jump index corresponding to each frequency component; Curve fitting is performed on the phase difference sequence corresponding to each frequency component to obtain the phase difference fitting curve for each frequency component. Calculate the absolute value of the derivative of each sequence point in the phase difference sequence corresponding to each frequency component in the corresponding phase difference fitting curve, and obtain the derivative values ​​of multiple sequence points corresponding to each frequency component. The average of the derivative values ​​of multiple sequence points corresponding to each frequency component is determined as the phase drift index corresponding to each frequency component. Based on the phase drift index and phase jump index corresponding to each frequency component, the phase change index corresponding to each frequency component is obtained.

9. The broadband adaptive filtering and reverse acoustic wave synthesis method for intelligent noise reduction equipment according to claim 1, characterized in that, The analysis of the amplitude variation of each product function component yields the human voice interference index for each product function component, including: Each product function component is decomposed into multiple signal segments corresponding to each product function component. Clustering is performed on multiple signal segments corresponding to each product function component to obtain multiple clusters corresponding to each product function component; Using the mean magnitude of the clusters as a comparison index, the multiple clusters corresponding to each product function component are sorted in descending order to obtain the cluster sequence corresponding to each product function component; In the cluster sequence corresponding to each product function component, the difference between the mean amplitudes of two adjacent clusters is calculated to obtain multiple cluster amplitude differences corresponding to each product function component, wherein the cluster amplitude difference is the difference between the mean amplitude of the corresponding cluster and its preceding cluster. Among the multiple cluster amplitude differences corresponding to each product function component, the cluster corresponding to the largest cluster amplitude difference is determined as the critical cluster corresponding to each product function component, and the multiple clusters located after the critical cluster in the corresponding cluster sequence are determined as the multiple strong signal clusters corresponding to each product function component. The amplitude changes among multiple strong signal clusters corresponding to each product function component are analyzed to obtain the human voice interference index for each product function component.

10. The broadband adaptive filtering and reverse acoustic wave synthesis method for intelligent noise reduction equipment according to claim 9, characterized in that, The analysis of amplitude changes among multiple strong signal clusters corresponding to each product function component, to obtain the human voice interference index for each product function component, includes: In each strong signal cluster corresponding to each product function component, the degree of cluster difference between each strong signal cluster and its adjacent next strong signal cluster is analyzed to obtain the cluster difference value corresponding to each product function component. In each strong signal cluster corresponding to each product function component, the difference between the average cluster amplitude of each strong signal cluster and its adjacent next strong signal cluster is calculated to obtain multiple target amplitude differences corresponding to each product function component. Calculate the ratio of the target magnitude difference to the corresponding cluster difference value for each component of the product function to obtain multiple cluster interference indices for each component of the product function. Calculate the mean of multiple cluster interference indices corresponding to each product function component to obtain the human voice interference index for each product function component.

Images