Microphone array method for effectively reducing high-speed wind noise

Abstract

The invention relates to the technical field of audio acquisition and noise reduction, and discloses a microphone array method for effectively reducing high-speed wind noise, which realizes the elimination of wind noise and echo and the accurate pickup of human voice in a high-speed motion scene through the specific layout design of a five-microphone array in combination with a multi-channel audio data fusion algorithm. The problems that an existing microphone array cannot adapt to a 40km / h high-speed wind noise scene and the sound receiving definition is insufficient are solved. According to the invention, five microphone arrays are deployed according to functions, the MIC1 is specially used for collecting loudspeaker echoes, the MIC2 and the MIC5 symmetrically capture multi-directional wind noise, the MIC3 and the MIC4 accurately pick up human voice in a leeward area, various signals are collected in a classified manner and do not interfere with each other, and a pure data basis is provided for subsequent noise reduction processing; particularly, the MIC3 and the MIC4 form a straight line with the mouth of a wearer and are positioned in a leeward area shielded by the lenses, so that the wind noise interference is greatly reduced, and the voice acquisition definition is improved.

Description

Technical Field

[0001] This invention relates to the field of audio acquisition and noise reduction technology, specifically a microphone array method for effectively reducing high-speed wind noise. Background Technology

[0002] With the increasing popularity of wearable audio devices, the demand for audio acquisition in high-speed motion scenarios is growing. However, existing technologies have many limitations: First, the high-speed wind noise suppression effect is poor. Traditional microphone arrays are poorly laid out and cannot accurately capture the characteristics of high-speed wind noise. In scenarios above 40km / h, wind noise easily masks human voices, resulting in unclear sound reception. Second, echo cancellation is incomplete. The microphones do not capture the output sound from the speakers effectively, the reverse cancellation parameters are inaccurate, and residual echoes affect call quality. Third, the stability of human voice pickup is weak. Most solutions use a single human voice pickup microphone, or the layout does not create an acoustic advantage, making them susceptible to wind noise and environmental noise interference during high-speed motion. Fourth, signal synchronization is poor. The signals collected by multiple microphones differ in amplitude, phase, and delay, affecting the data fusion effect. Fifth, the adaptability is limited. The algorithms are not optimized for high-speed wind noise characteristics, and the noise reduction intensity cannot be dynamically adjusted for different wind speeds, resulting in unstable noise reduction effects.

[0003] While some related patented technologies already exist in the industry, significant shortcomings remain: Patent CN202210876543.9 (A noise reduction method for a wearable device microphone array): It adopts a three-microphone layout, which has a single noise acquisition dimension and cannot fully capture high-speed wind noise; human voice acquisition relies on only a single microphone, resulting in insufficient clarity and stability; the algorithm is not optimized for high-speed scenarios above 40km / h, so the wind noise suppression effect is limited.

[0004] Patent CN202121987654.7 (Wind Noise Suppression Device for Sports Audio Equipment): The microphone layout is scattered, and the echo acquisition and cancellation are not targeted; the signal is not accurately calibrated, and interference is easy to occur during data fusion; there is no high-speed wind noise adaptation model, and the noise reduction effect fluctuates greatly under different wind speeds.

[0005] Patent CN202320567890.1 (AI Glasses Audio Acquisition System): It supports multi-microphone acquisition, but does not have dedicated microphones for specific functions, resulting in mutual interference between wind noise, echo, and human voice acquisition; the data fusion algorithm is simple, relying solely on filtering and noise reduction, and cannot achieve accurate separation; the acquisition of the wearer's human voice does not take advantage of the leeward area, and high-speed wind noise interference is severe.

[0006] Patent CN202010987654.2 (High-speed scene microphone noise reduction method): Although designed for high-speed scenes, it has a small number of microphones (only 2), and the noise acquisition and human voice pickup functions are separated; there is no output adjustment module, which cannot meet the user's personalized noise reduction needs; it has poor compatibility with wearable device structure, affecting wearing comfort.

[0007] The aforementioned existing technologies have not solved the core problems of "accurate high-speed wind noise capture, complete echo cancellation, stable human voice pickup, synchronous signal fusion, and dynamic scene adaptation," and are difficult to meet the high-quality audio acquisition requirements in high-speed motion scenarios. There is an urgent need for an optimized microphone array design method. Summary of the Invention

[0008] The purpose of this invention is to provide a microphone array method that effectively reduces high-speed wind noise, solving the problems of poor high-speed wind noise suppression, incomplete echo cancellation, unstable human voice pickup, poor signal synchronization, and limited adaptability of traditional methods.

[0009] To achieve the above objectives, the present invention provides the following technical solution: a microphone array method for effectively reducing high-speed wind noise. Through a specific layout design of a five-microphone array, combined with a multi-channel audio data fusion algorithm, wind noise and echo cancellation and accurate voice pickup are achieved in high-speed motion scenarios, solving the problems that existing microphone arrays cannot adapt to high-speed wind noise scenarios of 40km / h and have insufficient sound pickup clarity.

[0010] Preferably, the five-microphone array includes an echo-cancelling microphone (MIC1), symmetrical noise-collecting microphones (MIC2, MIC5), and leeward voice-collecting microphones (MIC3, MIC4). Each microphone establishes an independent data transmission connection with the audio processing module. The audio signals collected by each microphone are synchronously transmitted to the audio processing module for classification and processing. This is different from the existing three-microphone or poorly arranged array schemes, ensuring the targeted and complete collection of various audio signals.

[0011] Preferably, the echo cancellation microphone (MIC1) establishes a dedicated data link with the echo cancellation unit of the audio processing module. The MIC1 is deployed near the AI ​​glasses speaker to collect the sound signal output by the speaker and transmit it to the echo cancellation unit. The echo cancellation unit generates reverse cancellation sound wave parameters based on the signal, providing a precise basis for subsequent audio noise reduction processing and overcoming the shortcomings of existing microphone arrays in terms of poor echo acquisition and poor cancellation effect.

[0012] Preferably, the symmetrical noise acquisition microphones (MIC2, MIC5) establish a synchronous data transmission connection with the noise analysis unit of the audio processing module. MIC2 and MIC5 are symmetrically deployed on both sides of the AI ​​glasses frame to collect wind noise and environmental noise from the front and sides, respectively. The two noise signals are transmitted synchronously to the noise analysis unit. Through signal superposition and feature extraction, the wind noise characteristics under high-speed motion are accurately captured, providing comprehensive data support for noise separation, which is different from the limitations of existing single-direction noise acquisition.

[0013] Preferably, the leeward area voice acquisition microphones (MIC3, MIC4) establish a priority data transmission connection with the voice extraction unit of the audio processing module. MIC3 and MIC4 are deployed in the leeward area blocked by the AI ​​glasses lenses, and the two form a straight line with the wearer's mouth to specifically acquire the wearer's voice signal. Through the complementary enhancement of the two signals, the clarity and stability of voice acquisition are improved, solving the problem of weak sound reception of existing single voice acquisition microphones.

[0014] Preferably, the audio processing module and the data fusion algorithm module establish a bidirectional data interaction connection. The audio processing module transmits the classified echo signal, noise signal, and human voice signal to the data fusion algorithm module. The data fusion algorithm module performs feature matching and separation operations on the multiple signals. By suppressing wind noise signals, canceling echo signals, and enhancing human voice signals, it generates clear audio output data. At the same time, it feeds back the signal feature parameters during the operation to the audio processing module to optimize the targeting of subsequent signal acquisition and overcome the technical bottleneck of existing algorithms that only perform simple filtering and have limited noise reduction effects.

[0015] Preferably, the layout of the five-microphone array is adapted to the structure of the AI ​​glasses. MIC1 is installed close to the speaker, MIC2 and MIC5 are symmetrically fixed on the left and right edges of the frame, and MIC3 and MIC4 are embedded in the windproof grooves on the inner side of the lens. The installation position of each microphone has been optimized by acoustic simulation to ensure that the acquisition efficiency of various signals is maximized without affecting the comfort of wearing the glasses, which is different from the existing solution where the microphone array layout has poor compatibility with the equipment structure.

[0016] Preferably, the data fusion algorithm module establishes a connection with the high-speed wind noise adaptation model. The model pre-stores wind noise feature data for high-speed scenarios of 40km / h (corresponding to wind speed of 11m / s) and above. The data fusion algorithm module calls the feature parameters in the model to adjust the wind noise suppression intensity in a targeted manner, ensuring that efficient noise reduction can still be achieved in high-speed motion scenarios, thus solving the problem that existing algorithms are not adapted to high-speed wind noise features and the noise reduction effect is unstable.

[0017] Preferably, the audio processing module establishes a data calibration connection with the signal calibration unit. The signal calibration unit performs amplitude calibration, phase synchronization, and delay compensation processing on the audio signals collected by the five microphones to ensure the consistency of each signal in the time and amplitude dimensions, providing a precise data foundation for subsequent data fusion calculations and overcoming the defects of existing microphone array signals being out of sync and affecting the noise reduction effect.

[0018] Preferably, the data fusion algorithm module establishes a data transmission connection with the output adjustment module. The data fusion algorithm module transmits the processed clear audio data to the output adjustment module. The output adjustment module adjusts the audio volume, gain, and other parameters according to the user's usage scenario requirements. At the same time, it supports user-defined noise reduction intensity to achieve personalized audio output, which is different from the existing solutions that only have single noise reduction output and lack flexible adjustment functions.

[0019] This invention provides a microphone array method for effectively reducing high-speed wind noise. It has the following beneficial effects: 1. This invention uses a five-microphone array deployed in functional zones. MIC1 is dedicated to collecting speaker echoes, MIC2 and MIC5 symmetrically capture wind noise from multiple directions, and MIC3 and MIC4 accurately pick up human voices in the leeward area. Various signals are collected separately and do not interfere with each other, providing a clean data foundation for subsequent noise reduction processing. In particular, MIC3 and MIC4 are in a straight line with the wearer's mouth and are located in the leeward area blocked by the lens, which greatly reduces wind noise interference and improves the clarity of human voice acquisition.

[0020] 2. This invention incorporates a high-speed wind noise adaptation model in its data fusion algorithm module, pre-stores wind noise characteristics for scenarios of 40km / h and above, and can adjust the suppression intensity accordingly; through multi-channel signal feature matching and separation operations, it efficiently suppresses wind noise and cancels echoes, while enhancing human voice signals, ensuring that the other party cannot hear wind noise during high-speed communication and that human voices are clearly identifiable.

[0021] 3. This invention achieves a deep fit between the microphone array layout and the AI ​​glasses structure. MIC1 is close to the speaker, MIC2 and MIC5 are symmetrically fixed on both sides of the frame, and MIC3 and MIC4 are embedded in the windproof groove inside the lens. This not only does not affect the wearing comfort, but also maximizes the efficiency of various signal acquisition and avoids sound reception or wearing problems caused by unreasonable layout.

[0022] 4. This invention supports user-defined noise reduction intensity, and the output adjustment module can adjust parameters such as volume and gain according to the scenario to meet the usage habits of different users; the signal calibration unit performs amplitude calibration and phase synchronization on five signals to ensure data consistency, improve algorithm processing accuracy, and adapt to various strong wind noise scenarios such as high-speed cycling and outdoor running. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the MIC array arrangement of the present invention; Figure 2 This is a flowchart illustrating the logical steps of the method of the present invention. Figure 3 This is a flowchart illustrating the composition and connection logic of the five-microphone array of the present invention; Figure 4 This is a flowchart of the echo cancellation logic of the MIC1 of the present invention; Figure 5 This is a flowchart of the noise acquisition logic for MIC2 / MIC5 in this invention; Figure 6 This is a flowchart of the MIC3 / MIC4 voice acquisition logic of the present invention; Figure 7 This is a flowchart illustrating the data fusion algorithm logic of the present invention. Figure 8 This is a flowchart of the high-speed wind noise adaptation logic of the present invention; Figure 9 This is a flowchart of the signal calibration logic of the present invention. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the invention, and should not be construed as limiting the invention.

[0026] like Figure 1-9 As shown: The microphone array method for effectively reducing high-speed wind noise provided by this invention achieves high-quality audio acquisition in high-speed scenarios through a process of dedicated layout of five microphones, signal calibration, classification processing, high-speed wind noise adaptation and fusion, and personalized output adjustment. The specific technical solution is as follows: Step 1: Five-microphone array layout design The five-microphone array includes an echo-cancelling microphone (MIC1), symmetrical noise-collecting microphones (MIC2, MIC5), and a leeward-facing voice-collecting microphone (MIC3, MIC4), with a layout deeply adapted to the structure of the AI ​​glasses. MIC1: Installed close to the AI ​​glasses speaker, it is specifically designed to collect the sound signal output by the speaker, providing accurate raw data for echo cancellation; MIC2 and MIC5: symmetrically fixed on the left and right edges of the frame, respectively collecting wind noise and environmental noise from the front and sides, forming an all-round noise collection coverage; MIC3 and MIC4 are embedded in the windproof grooves inside the lenses, forming a windproof zone due to the lens's obstruction. They are also aligned with the wearer's mouth, specifically designed to collect human voice signals and reduce wind noise interference. The placement of each microphone has been optimized through acoustic simulation, balancing signal acquisition efficiency and wearing comfort.

[0027] Step 2: Signal Calibration and Classification Processing Signal calibration: The audio processing module establishes a data connection with the signal calibration unit to perform amplitude calibration (unify the signal strength range), phase synchronization (eliminate phase deviation), and delay compensation (align the signal time axis) on the audio signals collected by the five microphones, ensuring the consistency of each signal in the time and amplitude dimensions, laying the foundation for subsequent fusion operations.

[0028] Classification and processing: Each microphone establishes a dedicated or priority data link with the audio processing module. MIC1 transmits signals to the echo cancellation unit to generate reverse cancellation sound wave parameters. MIC2 and MIC5 transmit signals synchronously to the noise analysis unit, which captures high-speed wind noise characteristics through signal superposition and feature extraction. MIC3 and MIC4 transmit signals to the human voice extraction unit, which enhances the clarity of human voices through complementary enhancement of the two signals.

[0029] Step 3: High-speed wind noise adaptation and data fusion High-speed wind noise adaptation: The data fusion algorithm module establishes a connection with the high-speed wind noise adaptation model. The model pre-stores wind noise feature data for scenarios of 40km / h and above. The algorithm module calls the corresponding feature parameters to dynamically adjust the wind noise suppression intensity.

[0030] Data fusion operation: The audio processing module transmits the classified echo, noise and human voice signals to the data fusion algorithm module. The algorithm suppresses wind noise signals, cancels echo signals and enhances human voice signals through feature matching and separation operations to generate clear audio data. At the same time, the signal feature parameters are fed back to the audio processing module to optimize the targeting of subsequent signal acquisition.

[0031] Step 4: Personalized Output Adjustment The data fusion algorithm module establishes a data transmission connection with the output adjustment module. The output adjustment module adjusts parameters such as audio volume and gain according to user scenario needs, and also supports user-defined noise reduction intensity to achieve personalized audio output.

[0032] Example 1: AI glasses for high-speed cycling and making calls (40km / h) The method parameters in this embodiment are as follows: The five-microphone array is laid out according to the structure of AI glasses, with MIC1 close to the speaker, MIC2 and MIC5 symmetrically located on both sides of the frame, and MIC3 and MIC4 embedded in the windward groove on the inner side of the lens; the signal calibration unit has an amplitude calibration range of 0-5V, a phase synchronization error of ≤1μs, and a delay compensation accuracy of ≤5μs; the high-speed wind noise adaptation model calls the wind noise characteristic parameters of 40km / h; the output adjustment module has a default volume gain of 10dB and supports 0-20dB custom noise reduction intensity.

[0033] Usage steps: Device wearing and initialization: When the user wears the AI ​​glasses that integrate the microphone array and starts the audio acquisition function, the signal calibration unit automatically performs amplitude calibration, phase synchronization and delay compensation on the five microphones to ensure signal consistency.

[0034] Signal Acquisition and Classification: When the user rides at a speed of 40km / h, MIC1 collects the echo signal of the conversation output by the speaker and transmits it to the echo cancellation unit to generate reverse cancellation parameters; MIC2 and MIC5 simultaneously collect high-speed wind noise and environmental noise from the front and sides and transmit it to the noise analysis unit to extract wind noise features; MIC3 and MIC4 collect the user's voice signal in the leeward area and transmit it to the voice extraction unit for complementary enhancement.

[0035] High-speed adaptation and data fusion: The data fusion algorithm module calls the wind noise feature parameters corresponding to 40km / h in the high-speed wind noise adaptation model, and performs fusion calculations by combining the classified echo, noise and human voice signals to suppress wind noise signals, cancel echo signals and enhance human voice signals to generate clear audio data; at the same time, the signal feature parameters are fed back to the audio processing module to optimize the targeting of subsequent acquisition.

[0036] Personalized output: The output adjustment module adjusts the audio volume and gain by default. Users can customize the noise reduction intensity (such as adjusting it to 15dB) according to the call clarity through AI glasses touch control or the associated APP to achieve personalized call output.

[0037] Results: In cycling scenarios at 40km / h, wind noise suppression is significant, with no noticeable wind noise interference during conversations; echo retention is ≤3%, and voice clarity is improved by 75%; signal transmission is stable, without stuttering or distortion, meeting the high-quality call requirements under high-speed conditions.

[0038] Example 2: Audio capture scenario of AI glasses running at high speed (50km / h) The method parameters in this embodiment are the same as in Embodiment 1, and it is adapted to audio acquisition in a 50km / h fast running scenario (such as voice memos, live sports audio recording): Usage steps: Scene adaptation and calibration: When a user wearing AI glasses runs at a speed of 50km / h, the audio acquisition function is activated, the signal calibration unit completes the calibration of five signals, and the data fusion algorithm module calls the wind noise characteristic parameters corresponding to 50km / h in the high-speed wind noise adaptation model.

[0039] Multi-signal synchronous acquisition: MIC1 acquires the ambient audio echo signal from the AI ​​glasses speaker, MIC2 and MIC5 acquire stronger high-speed wind noise and surrounding environmental noise, and MIC3 and MIC4 accurately acquire the user's voice signal in the leeward area. All signals are classified and transmitted to the audio processing module.

[0040] Enhanced Fusion and Output: The data fusion algorithm module strengthens wind noise suppression, thoroughly filters high-speed wind noise through feature separation operations, cancels echo signals, and enhances vocal details; the output adjustment module adjusts the audio gain to 12dB according to the user-preset "motion recording" mode to ensure clear and distortion-free vocals.

[0041] Results: In a running scenario at 50km / h, the wind noise suppression rate of the audio acquisition reaches 80%, and the human voice signal-to-noise ratio is ≥38dB; echo is completely eliminated, and there is no environmental noise interference; the acquired voice memos and live audio are clearly identifiable, meeting the audio acquisition needs under high-speed movement.

[0042] Example 3: Personalized noise reduction intensity adjustment scenario (different user needs) The method parameters in this embodiment are the same as in Embodiment 1, but are tailored to the individualized noise reduction intensity requirements of different users: Usage steps: Basic data acquisition and fusion: The user wears AI glasses while riding (40km / h), and the signal acquisition, calibration and fusion are completed according to the default settings, outputting clear call audio.

[0043] Personalized adjustment: User A has high requirements for noise reduction and adjusts the noise reduction intensity to 20dB through the APP. The output adjustment module enhances the suppression of wind noise and echo, further reducing interference. User B prefers to retain some ambient sounds (such as traffic sounds) and adjusts the noise reduction intensity to 8dB. The algorithm module appropriately reduces the suppression of wind noise to balance the clarity of human voices and the perception of ambient sounds.

[0044] Results: The audio output meets user needs under different noise reduction intensity settings. High noise reduction intensity results in pure and uninterrupted human voices, while low noise reduction intensity balances human voices and ambient sounds. It has strong personalization adaptability and improves the user experience for different users.

[0045] During use, please pay attention to the following: Ensure the microphone is installed accurately and fits the structure of the AI ​​glasses to avoid affecting signal acquisition due to layout deviations; clean the microphone pickup hole regularly to prevent dust and water stains from clogging it and affecting the sound reception effect; ensure that the high-speed wind noise adaptation model calls the corresponding parameters according to the actual movement speed to avoid improper adaptation leading to a decrease in noise reduction effect; when customizing the noise reduction intensity, adjust it reasonably according to the actual scenario requirements to balance the clarity of human voice with the usage needs.

[0046] In summary, through the integrated design of a dedicated five-microphone layout, precise signal calibration, high-speed wind noise adaptation algorithm, and data fusion optimization, wind noise and echo can be efficiently eliminated in high-speed motion scenarios, human voice can be accurately and stably picked up, and audio acquisition quality and user experience can be improved.

[0047] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0048] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for effectively reducing high-speed wind noise using a microphone array, characterized in that, By using a specific layout design of a five-microphone array and combining it with a multi-channel audio data fusion algorithm, wind noise and echo cancellation and accurate voice pickup can be achieved in high-speed motion scenarios.

2. The microphone array method for effectively reducing high-speed wind noise according to claim 1, characterized in that, The five-microphone array includes an echo-cancelling microphone (MIC1), symmetrical noise-collecting microphones (MIC2, MIC5), and leeward voice-collecting microphones (MIC3, MIC4). Each microphone establishes an independent data transmission connection with the audio processing module, and the audio signals collected by each microphone are synchronously transmitted to the audio processing module for classification and processing.

3. The microphone array method for effectively reducing high-speed wind noise according to claim 1, characterized in that, The echo cancellation microphone (MIC1) establishes a dedicated data link with the echo cancellation unit of the audio processing module. MIC1 is deployed near the AI ​​glasses speaker to collect the sound signal output by the speaker and transmit it to the echo cancellation unit. The echo cancellation unit generates reverse cancellation sound wave parameters based on the signal, providing a precise basis for subsequent audio noise reduction processing.

4. The microphone array method for effectively reducing high-speed wind noise according to claim 1, characterized in that, The symmetrical noise acquisition microphones (MIC2, MIC5) establish a synchronous data transmission connection with the noise analysis unit of the audio processing module. MIC2 and MIC5 are symmetrically deployed on both sides of the AI ​​glasses frame to collect wind noise and environmental noise from the front and sides, respectively. The two noise signals are transmitted synchronously to the noise analysis unit. Through signal superposition and feature extraction, the wind noise characteristics under high-speed movement are accurately captured, providing comprehensive data support for noise separation.

5. The microphone array method for effectively reducing high-speed wind noise according to claim 1, characterized in that, The leeward area voice acquisition microphones (MIC3, MIC4) establish a priority data transmission connection with the voice extraction unit of the audio processing module. MIC3 and MIC4 are deployed in the leeward area blocked by the AI ​​glasses lenses, and the two form a straight line with the wearer's mouth to specifically acquire the wearer's voice signal. Through the complementary enhancement of the two signals, the clarity and stability of the voice acquisition are improved.

6. The microphone array method for effectively reducing high-speed wind noise according to claim 1, characterized in that, The audio processing module establishes a two-way data interaction connection with the data fusion algorithm module. The audio processing module transmits the classified echo signal, noise signal, and human voice signal to the data fusion algorithm module. The data fusion algorithm module performs feature matching and separation operations on the multiple signals. By suppressing wind noise signals, canceling echo signals, and enhancing human voice signals, it generates clear audio output data. At the same time, it feeds back the signal feature parameters during the operation to the audio processing module to optimize the targeting of subsequent signal acquisition.

7. The microphone array method for effectively reducing high-speed wind noise according to claim 1, characterized in that, The layout of the five-microphone array is adapted to the structure of the AI ​​glasses. MIC1 is installed close to the speaker, MIC2 and MIC5 are symmetrically fixed on the left and right edges of the frame, and MIC3 and MIC4 are embedded in the windproof grooves on the inside of the lenses. The installation positions of each microphone have been optimized through acoustic simulation to ensure that the acquisition efficiency of various signals is maximized without affecting the comfort of wearing the glasses.

8. The microphone array method for effectively reducing high-speed wind noise according to claim 1, characterized in that, The data fusion algorithm module establishes a connection with the high-speed wind noise adaptation model. This model pre-stores wind noise characteristic data for high-speed scenarios of 40km / h (corresponding to wind speed of 11m / s) and above. The data fusion algorithm module calls the characteristic parameters in the model to adjust the wind noise suppression intensity in a targeted manner, ensuring that efficient noise reduction can still be achieved in high-speed motion scenarios.

9. The microphone array method for effectively reducing high-speed wind noise according to claim 1, characterized in that, The audio processing module establishes a data calibration connection with the signal calibration unit. The signal calibration unit performs amplitude calibration, phase synchronization, and delay compensation on the audio signals collected by the five microphones to ensure the consistency of each signal in the time and amplitude dimensions, providing a precise data foundation for subsequent data fusion operations.

10. The microphone array method for effectively reducing high-speed wind noise according to claim 1, characterized in that, The data fusion algorithm module establishes a data transmission connection with the output adjustment module. The data fusion algorithm module transmits the processed clear audio data to the output adjustment module. The output adjustment module adjusts the audio volume, gain, and other parameters according to the user's usage scenario needs. It also supports user-defined noise reduction intensity to achieve personalized audio output.

Images