CN111833890A - Device and method for automatically detecting wearing state of helmet - Google Patents
Device and method for automatically detecting wearing state of helmet Download PDFInfo
- Publication number
- CN111833890A CN111833890A CN202010671306.5A CN202010671306A CN111833890A CN 111833890 A CN111833890 A CN 111833890A CN 202010671306 A CN202010671306 A CN 202010671306A CN 111833890 A CN111833890 A CN 111833890A
- Authority
- CN
- China
- Prior art keywords
- helmet
- wind noise
- ear
- microphone
- microphones
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 49
- 230000005236 sound signal Effects 0.000 claims abstract description 114
- 238000001228 spectrum Methods 0.000 claims abstract description 86
- 238000001514 detection method Methods 0.000 claims abstract description 42
- 238000012546 transfer Methods 0.000 claims abstract description 38
- 210000000883 ear external Anatomy 0.000 claims description 35
- 238000012545 processing Methods 0.000 claims description 14
- 230000007613 environmental effect Effects 0.000 claims description 13
- 238000005314 correlation function Methods 0.000 claims description 11
- 239000002131 composite material Substances 0.000 claims description 8
- 230000002194 synthesizing effect Effects 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 abstract description 9
- 230000000694 effects Effects 0.000 abstract description 6
- 230000001427 coherent effect Effects 0.000 abstract description 5
- 238000002955 isolation Methods 0.000 abstract description 3
- 230000006870 function Effects 0.000 description 32
- 238000010586 diagram Methods 0.000 description 16
- 238000007689 inspection Methods 0.000 description 6
- 230000003595 spectral effect Effects 0.000 description 3
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 238000009499 grossing Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000033228 biological regulation Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 210000005069 ears Anatomy 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007500 overflow downdraw method Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
- G10L19/02—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using spectral analysis, e.g. transform vocoders or subband vocoders
- G10L19/0212—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using spectral analysis, e.g. transform vocoders or subband vocoders using orthogonal transformation
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
- G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Signal Processing (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Circuit For Audible Band Transducer (AREA)
Abstract
The application relates to a device and a method for automatically detecting the wearing state of a helmet, wherein the method comprises the following steps: picking up sound signals of at least 2 microphones; respectively carrying out fast Fourier transform on the sound signals of the at least 2 microphones to obtain short-time spectrums of each frequency band; determining whether the current frame of the sound signal is wind noise according to the short-time spectrums of the frequency bands; and calculating the wind noise frame number in the sound signal, and judging the wearing condition of the helmet in the advancing state according to the wind noise frame number. The embodiment of the application provides a helmet wearing automatic detection method, based on the shielding effect of a helmet on wind noise, the isolation effect on external environment noise and the influence on the voice of a wearer, helmet wearing detection is carried out through coherent calculation and relative transfer function calculation, and experimental results show that the helmet wearing detection method is high in detection performance, accurate in detection result and capable of meeting application requirements.
Description
Technical Field
The application relates to the field of acoustic detection, in particular to a device and a method for automatically detecting the wearing state of a helmet.
Background
Motorcycles and electric vehicles are commonly used vehicles because of their small size, high speed, and agile mobility; however, most of the drivers generally lack the safety protection awareness, for example, some drivers do not wear safety protection equipment such as helmets according to regulations, which increases the safety risk of the drivers and also harms the life and property safety of others; whether the driver correctly wears the helmet in the driving state or not is automatically detected, and the method has important application value. The traditional method needs an additional helmet wearing sensor and is installed on the helmet, and meanwhile, a voice reminding device needs to be installed on the helmet, which increases the difficulty and cost of manufacturing the helmet.
In recent years, with the progress of wireless transmission technology and the miniaturization and miniaturization of chips, a true wireless stereo headset has begun to be popularized, and has advantages of lightness, binaural sound pickup, binaural reproduction, and the like, compared to a monaural conventional bluetooth headset. In driving a motor vehicle, hands-free operation is generally required, and communication using a true wireless stereo headset is becoming the mainstream.
Disclosure of Invention
The purpose of this application is to solve traditional helmet and wear detection device and method and need the problem that extra helmet wearing sensor increases helmet manufacturing difficulty and cost.
In order to achieve the above object, in a first aspect of the present application, there is provided a method for automatically detecting a wearing state of a helmet, including: picking up sound signals of at least 2 microphones; respectively carrying out fast Fourier transform on the sound signals of the at least 2 microphones to obtain short-time spectrums of each frequency band; determining whether the current frame of the sound signal is wind noise according to the short-time spectrums of the frequency bands; and calculating the wind noise frame number in the sound signal, and judging the wearing condition of the helmet in the advancing state according to the wind noise frame number.
In one possible embodiment, the method further comprises: calculating a cross-correlation function of sound signals among the microphones in the at least 2 microphones according to the short-time spectrums of the frequency bands; calculating a composite sound source incident angle estimation value among all microphones in the at least 2 microphones according to the cross-correlation function; and judging whether the current frame is the voice of the wearer according to the composite sound source incident angle estimated value, and judging wind noise or environmental noise if the current frame is not the voice of the wearer.
In one possible embodiment, the at least 2 microphones are binaural microphones; the determining whether the current frame of the sound signal is wind noise according to the short-time spectrums of the frequency bands comprises: calculating coherence between the double-ear external sound transmitters according to the short-time spectrums of the frequency bands, and judging whether the absolute value of the coherence meets the numerical condition that the current frame of the sound signal is the wind noise; synthesizing at least 2 frequency band short-time spectrums into a sub-frequency band short-time spectrum; estimating the incidence angle of the sound source for each sub-band short-time spectrum, and calculating the variance of the incidence angle; judging whether the incidence angle variance meets the numerical condition that the current frame of the sound signal is wind noise; and when both are in accordance with the numerical condition of wind noise, determining that the current frame of the sound signal is wind noise.
In one possible embodiment, the at least 2 microphones are an out-of-ear microphone and an in-ear feedback microphone; the determining whether the current frame of the sound signal is wind noise according to the short-time spectrums of the frequency bands comprises: calculating coherence between an ear external microphone and an ear internal feedback microphone according to the short-time spectrums of the frequency bands, and judging whether the absolute value of the coherence between the microphones meets the numerical condition that the current frame of the sound signal is wind noise; calculating a module value of a relative transfer function fusing at least 2 frequency band short-time spectrums, and judging whether the module value meets the numerical condition that the current frame of the sound signal is wind noise; and when both are in accordance with the numerical condition of wind noise, determining that the current frame of the sound signal is wind noise.
In one possible embodiment, the calculating a wind noise frame number in the sound signal and determining the wearing condition of the helmet in the traveling state according to the wind noise frame number includes: counting the number of wind noise frames in the sound signal within a period of time; judging whether the wind noise frame number meets the numerical condition that the helmet is not worn in the advancing state; if yes, the helmet is not worn in the traveling state at present.
In one possible embodiment, the at least 2 microphones are a binaural external microphone and an in-ear feedback microphone; the method further comprises the following steps: calculating coherence between the double-ear external sound transmitters according to the short-time spectrums of the frequency bands, and judging whether the absolute value of the coherence meets the numerical condition that the current frame of the sound signal is the wind noise; synthesizing at least 2 frequency band short-time spectrums into a sub-frequency band short-time spectrum; estimating the incidence angle of the sound source for each sub-band short-time spectrum, and calculating the variance of the incidence angle; judging whether the incidence angle variance meets the numerical condition that the current frame of the sound signal is wind noise; when both are in accordance with the numerical condition of wind noise, determining that the current frame of the sound signal is wind noise; counting the number of wind noise frames of the sound signal within a period of time under the condition of the double-ear external sound transmitter to obtain a first number of wind noise frames; calculating coherence between an out-of-ear microphone and an in-ear feedback microphone according to the short-time spectrums of the frequency bands, and judging whether the absolute value of the coherence between the microphones meets the numerical condition that the current frame of the sound signal is wind noise; calculating a module value of a relative transfer function fusing at least 2 frequency band short-time spectrum frequency bands, and judging whether the module value meets the numerical condition that the current frame of the sound signal is wind noise; when both of the two values meet the numerical condition that the current frame of the sound signal is wind noise, determining the current frame of the sound signal to be wind noise; counting the number of wind noise frames of the sound signal within a period of time under the conditions of an out-of-ear microphone and an in-ear feedback microphone to obtain a second number of wind noise frames; and adding the values of the first wind noise frame number and the second wind noise frame number, and judging whether the helmet is worn in the current advancing state according to the addition result.
In one possible embodiment, the at least 2 microphones are binaural microphones, the method further comprising the step of detecting the wearing of the helmet in the non-travelling state: calculating the absolute value of the coherence between the double-ear external sound transmitters according to the short-time spectrums of the frequency bands; accumulating absolute values of coherence of the microphones over a full frequency band of the sound signal; and judging whether the accumulated result meets the numerical condition of wearing the helmet or not, and judging whether the helmet is worn in the current non-advancing state or not.
In one possible embodiment, the at least 2 microphones are an out-of-ear microphone and an in-ear feedback microphone, the method further comprising the step of detecting the wearing of the helmet in a non-travelling state: calculating the module value of the relative transfer function from the ear external microphone to the ear internal feedback microphone according to the short-time spectrums of the frequency bands; obtaining typical values under the condition of not wearing the helmet, wherein the first typical value is a relative transfer function module value of environmental noise, and the second typical value is a relative transfer function module value of the voice of a wearer; if the current frame is the voice of the wearer, judging whether the module value of the relative transfer function and the second typical value meet the numerical condition of wearing the helmet or not, and judging whether the helmet is worn in the current non-advancing state or not; and if the current frame is the environmental noise, judging whether the module value of the relative transfer function and the first typical value meet the numerical condition of wearing the helmet or not, and judging whether the helmet is worn in the current non-advancing state or not.
In one possible embodiment, the method further comprises determining from the GPS signal whether the helmet is in a travel state at the time of wear detection.
In a second aspect of the present application, there is provided an apparatus for automatically detecting a wearing state of a helmet, the apparatus comprising a microphone and a digital signal processing module on an earphone; the microphone is used for picking up sound signals; the number of the microphones is at least 2; the digital signal processing module is used for respectively carrying out fast Fourier transform on the sound signals of the at least 2 microphones to obtain short-time spectrums of each frequency band; determining whether the current frame of the sound signal is wind noise according to the short-time spectrums of the frequency bands; and calculating the number of wind noise frames in the sound signal, and judging the wearing condition of the helmet in the advancing state.
In one possible embodiment, the microphones include an out-of-ear microphone and an in-ear feedback microphone.
In a possible embodiment, the digital signal processing module is further configured to calculate an absolute value of coherence between binaural external speakers according to the short-time spectrums of the respective frequency bands, and determine whether a numerical condition for wearing the helmet in a non-traveling state is met.
In a possible implementation manner, the digital signal processing module is further configured to calculate a module value of a relative transfer function from the ear-external microphone to the ear-internal feedback microphone according to the short-time spectrum of each frequency band, and determine whether a numerical condition for wearing the helmet in a non-traveling state is met.
In one possible embodiment, the device further comprises a GPS locator for determining whether the headgear is worn in a travelling state.
Compared with the prior art, the application has the beneficial effects that:
the embodiment of the application provides a helmet wearing automatic detection method, based on the shielding effect of a helmet on wind noise, the isolation effect on external environment noise and the influence on the voice of a wearer, helmet wearing detection is carried out through coherent calculation and relative transfer function calculation, and experimental results show that the helmet wearing detection method is high in detection performance, accurate in detection result and capable of meeting application requirements.
The embodiment of the application provides a helmet wearing state automated inspection's device, the condition of wearing of carrying on the helmet when motorcycle and electric motor car driver wear the earphone detects, utilizes the microphone signal on the earphone to carry out the automated inspection of helmet wearing state, compares in traditional helmet wearing detection device, and the device need not install on the helmet, and only uses the earphone of daily use can realize the automated inspection of helmet wearing state, need not extra sensor and facility.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments disclosed in the present application, the drawings required to be used in the description of the embodiments will be briefly introduced below, and it is apparent that the drawings in the following description are only embodiments disclosed in the present application, and it is obvious for those skilled in the art that other drawings can be obtained based on the drawings without inventive efforts.
FIG. 1 is a schematic diagram of a microphone array configuration commonly used in earphones, wherein (1a) a binaural external microphone; (1b) a binaural external microphone and an in-ear feedback microphone; (1c) an out-of-ear microphone and an in-ear feedback microphone;
FIG. 2 is a conventional motorcycle or electric vehicle helmet;
fig. 3 is a frequency spectrum diagram of sound signals picked up by an ear microphone and an ear feedback microphone in a helmet-mounted state according to an embodiment of the present application; wherein:
FIG. 3a is a diagram showing the spectrum of the sound signal of the first extra-aural sound transmitter in the helmet-mounted state;
FIG. 3b is a diagram showing the spectrum of the sound signal of the second extra-aural sound transmitter in the helmet-mounted state;
FIG. 3c is a spectrum diagram of the sound signal of the in-ear feedback microphone in a helmet-mounted state;
fig. 4 is a frequency spectrum diagram of sound signals picked up by an ear microphone and an ear feedback microphone in a helmet-free state according to an embodiment of the present application; wherein:
FIG. 4a is a spectrum diagram of a first extra-aural sound transmitter in a helmet-free state;
FIG. 4b is a spectrum diagram of a second extra-aural sound transmitter in a helmet-free state;
FIG. 4c is a spectrum diagram of an in-ear feedback microphone in the helmet-free state;
fig. 5 is a coherence and an energy ratio of sound signals picked up by a microphone in a helmet-mounted state according to an embodiment of the present application; wherein:
FIG. 5a is a diagram showing the coherence spectrum of the sound signals of the first and second external-ear microphones when the helmet is worn;
FIG. 5b is a graph showing the energy ratio spectrum of the sound signals of the first and second external-ear microphones when the helmet is worn;
FIG. 5c is a signal coherence spectrum diagram of the binaural external microphone and the ear feedback microphone in a helmet wearing state;
FIG. 5d is a signal energy ratio spectrum diagram of the binaural external microphone and the in-ear feedback microphone in a helmet wearing state;
fig. 6 shows coherence and energy ratio of sound signals picked up by a microphone in a helmet-free state according to an embodiment of the present application; wherein:
fig. 6a shows the coherence of the sound signals of the first and second extra-aural microphones when the helmet is not worn;
FIG. 6b is a graph showing the sound signal energy ratio of the first and second extra-aural speakers when the helmet is not worn;
FIG. 6c shows the signal coherence of the binaural external microphone and the ear feedback microphone in the helmet-free state;
FIG. 6d is a graph of the energy ratio of two first out-of-the-ear microphone sound signals to the in-the-ear feedback microphone signal in a non-helmet wearing state;
fig. 7 is a schematic diagram of a method for automatically detecting a wearing state of a helmet according to an embodiment of the present application;
fig. 8 is a flowchart of a method for automatically detecting a wearing state of a helmet according to an embodiment of the present application, for determining a wearing state of a helmet in a traveling state in a case of a binaural external microphone;
fig. 9 is a flowchart of a method for automatically detecting a wearing state of a helmet according to an embodiment of the present application, for determining a wearing state of a helmet in a traveling state with an out-of-ear microphone and an in-ear feedback microphone;
fig. 10 is a flowchart of a method for automatically detecting a wearing state of a helmet according to an embodiment of the present application, for determining a wearing state of a helmet in a traveling state under a condition of a binaural external microphone and an in-ear feedback microphone;
fig. 11 is a flowchart of a method for automatically detecting a wearing state of a helmet according to an embodiment of the present application, for determining a wearing condition of a helmet in a non-traveling state in the case of a binaural external microphone;
fig. 12 is a flowchart of a method for automatically detecting a wearing state of a helmet according to an embodiment of the present application, for determining a wearing state of the helmet in a non-traveling state with one out-of-ear microphone and one in-ear feedback microphone.
Detailed Description
The technical solution of the present application is further described in detail by the accompanying drawings and examples.
The device for automatically detecting the wearing state of the helmet provided by the embodiment of the application can be a single-side earphone which is provided with an in-ear feedback microphone and at least one out-of-ear microphone, or the single-side earphone is provided with two out-of-ear microphones. Typical single-sided headphones are, for example, active noise reduction headphones having an external binaural microphone and an in-ear feedback microphone.
The device for automatically detecting the wearing state of the helmet provided by the embodiment of the application can also be a binaural earphone, typically a true wireless stereo earphone, for example, if the true wireless stereo earphone has a binaural external microphone, the binaural external microphone can be respectively arranged outside the left and right earphones; if the true wireless stereo headset is provided with an in-ear feedback microphone, the in-ear feedback microphone can be arranged in a side headset; if the true wireless stereo headset has only one extra-aural microphone and one in-aural feedback microphone, it is possible to arrange the two microphones on the same side or on different sides.
Fig. 1 is a schematic view of a microphone array configuration commonly used in earphones, as shown in fig. 1, wherein (1a) is a binaural external microphone; (1b) the two ear external microphones and an ear internal feedback microphone; (1c) an out-of-ear microphone and an in-ear feedback microphone.
FIG. 2 is a conventional motorcycle or electric vehicle helmet; as shown in fig. 2, the helmet determines whether the helmet is worn in a traveling state by a sensor, such as a GPS locator on the helmet; the digital signal processing module can also process the picked sound signals of the left and right ear microphones to carry out joint judgment.
In practical application, the detection significance of the helmet wearing state of the advancing state is larger, and the helmet wearing state of the advancing state can effectively protect the safety of a driver, so that the method solves the problem of detecting the helmet wearing state of the motorcycle and the electric vehicle driver in the advancing state on one hand, provides a method for detecting according to wind noise characteristics, and solves the problem of detecting the helmet wearing state of the motorcycle and the electric vehicle driver in the non-advancing state on the other hand. Although the detection of the wearing state of the helmet in the non-traveling state is not important from the safety perspective, the detection of the wearing state of the helmet in the non-traveling state is also significant from the perspective of improving the voice communication quality of the earphone wearer, and can provide an important basis for the signal processing of the earphone microphone.
The embodiment of the application provides a device for automatically detecting the wearing state of a helmet, which comprises a microphone and a digital signal processing module on an earphone. The portable wearable device earphone can be used as a device for detecting the wearing state of the helmet, and the automatic detection of the wearing state of the helmet is realized through a microphone and a digital signal processing module which are arranged in the earphone.
Specifically, the number of microphones is at least 2, and the microphones pick up voice signals; the digital signal processing module respectively carries out fast Fourier transform on sound signals of at least 2 microphones to obtain short-time spectrums of each frequency band; determining whether the current frame of the sound signal is wind noise according to the short-time spectrums of the frequency bands; and calculating the number of wind noise frames in the sound signal so as to judge the wearing condition of the helmet in the traveling state.
In a non-advancing state, the digital signal processing module also calculates the absolute value of coherence between the double-ear external sound transmitters according to the short-time spectrums of all the frequency bands, and judges whether the numerical condition of wearing the helmet is met; or the digital signal processing module calculates the module value of the relative transfer function between the ear external microphone and the ear internal feedback microphone according to the short-time spectrums of the frequency bands, and judges whether the numerical condition of wearing the helmet is met or not, so that the wearing condition of the helmet in a non-advancing state is judged.
The embodiment of the application provides a helmet wearing state automated inspection's device, carries out the helmet and wears the condition when the earphone is worn to motorcycle and electric motor car driver and detects, utilizes the microphone signal on the earphone to carry out the automated inspection of helmet wearing state, compares in traditional helmet wearing detection device, and the device need not install on the helmet, and only uses the earphone of daily use can realize the automated inspection of helmet wearing state, need not extra sensor and facility.
The detection basis of the automatic helmet wearing state detection device is analyzed.
FIG. 3 is a spectrum diagram of sound signals picked up by an ear microphone and an ear feedback microphone while the helmet is worn; wherein (3a) is a sound signal spectrogram of a first ear-external microphone under the helmet wearing state, (3b) is a sound signal spectrogram of a second ear-external microphone under the helmet wearing state, and (3c) is a sound signal spectrogram of an ear-internal feedback microphone under the helmet wearing state.
FIG. 4 is a graph of the frequency spectra of the sound signals picked up by the extra-aural and in-ear feedback microphones when the helmet is not worn; wherein (4a) is a spectrogram of the first ear-external microphone in a helmet-free state, (4b) is a spectrogram of the second ear-external microphone in a helmet-free state, and (4c) is a spectrogram of the ear-internal feedback microphone in a helmet-free state.
Analyzing the spectral characteristics of the sound signal comparing the spectrograms shown in fig. 3 and 4 can obtain: when the helmet is worn, the ear external microphone and the ear internal feedback microphone are not influenced by wind noise; when the helmet is not worn, the out-of-ear microphone is greatly affected by wind noise, while the in-ear feedback microphone is hardly affected by wind noise.
FIG. 5 is a diagram showing coherence and energy ratio of sound signals picked up by a microphone in a helmet-mounted state; wherein (5a) is the coherence of the sound signal of the first ear external sound transmitter and the sound signal of the second ear external sound transmitter under the state of wearing the helmet, (5b) is the energy ratio of the sound signal of the first ear external sound transmitter and the sound signal of the second ear external sound transmitter under the state of wearing the helmet, (5c) is the signal coherence of the ear external sound transmitter and the ear internal feedback microphone under the state of wearing the helmet, and (5d) is the signal coherence and the energy ratio of the ear external sound transmitter and the ear internal feedback microphone under the state of wearing the helmet.
FIG. 6 is a diagram showing coherence and energy ratio of sound signals picked up by a microphone in a state where a helmet is not worn; wherein (6a) the first and second out-of-the-ear microphone sound signals are coherent in the helmet-free state; (6b) the energy ratio of the sound signal of the first ear external microphone to the sound signal of the second ear external microphone is set under the state that the helmet is not worn; (6c) the signal coherence of the binaural external microphone sound signal and the in-ear feedback microphone signal in the helmet-free state, and (6d) the energy ratio of the two first in-ear external microphone sound signals and the in-ear feedback microphone signal in the helmet-free state.
Comparing the coherence and energy ratio spectral characteristics shown in fig. 5 and 6 yields: because the ear cavity resonance causes the ear feedback microphone to have larger energy at low frequency, the helmet wearing state and the helmet wearing state show obvious difference in coherence and energy ratio. Under the state of not wearing the helmet, no matter between the ear external microphones or between the ear external microphones and the ear internal feedback microphones, the signal coherence is lower, and the low frequency performance is more obvious; when the helmet is worn, the coherence between the ear external microphones or between the ear external microphones and the ear internal feedback microphones is high. Because wind noise is generally high in energy, when the helmet is not worn, the low-frequency energy ratio between the ear external microphone and the ear internal feedback microphone is remarkably increased; conversely, when the helmet is worn, the low frequency energy ratio between the out-of-ear microphone and the in-ear feedback microphone may decrease.
As described above, the helmet wearing state, such as the difference in the influence of wind noise on the ear-external microphone and the ear-internal feedback microphone, the difference in the influence of environmental noise on the ear-external microphone and the ear-internal feedback microphone, and the influence of the closed space on the estimation of the incident angle of the sound source, can be detected from the difference in the sound signals picked up by the ear-external microphone and the ear-internal feedback microphone in the helmet wearing state and the helmet non-wearing state.
In accordance with the above analysis, another embodiment of the present application proposes a method for automatically detecting a wearing state of a helmet, as shown in fig. 7, the method comprising: s1, picking up sound signals of at least 2 microphones; s2, respectively carrying out fast Fourier transform on the sound signals of the at least 2 microphones to obtain short-time spectrums of each frequency band; s3, determining whether the current frame of the sound signal is wind noise according to the short-time spectrums of the frequency bands; and S4, calculating the wind noise frame number in the sound signal, and judging the wearing condition of the helmet in the advancing state according to the wind noise frame number.
Specifically, in executing S1, sound signals are picked up by at least 2 microphones, and acoustic models are established for the 2 sound signals.
For the out-of-ear microphone and the in-ear feedback microphone, the sound signal x received by the ith out-of-ear microphone at the moment n is seti(n) is:
xi(n)＝si(n)+di(n)+dw,i(n) (1)
wherein s isi(n) the speech signal of the wearer received by the ith extra-aural microphone at time n, di(n) is the ith ear at time nAmbient noise received by an external microphone, dw,i(n) is the wind noise received by the ith extra-aural sound transmitter at the moment n; and i is 1,2 and … M, wherein M is the number of the out-of-ear microphones.
n time ith ear feedback sound signal received by microphone
Comprises the following steps:
wherein the content of the first and second substances,
and
the voice signal, the environmental noise and the wind noise of the wearer received by the ith in-ear feedback microphone at the moment n are respectively; i is 1,2, N, where N is the number of in-ear feedback microphones, and N is 1 in this embodiment.
In executing S2, taking the headphone configuration composed of the single-sided headphone and the binaural external microphone as an example, the signals received by the binaural external microphone are x1(n) and x2(n), the frequency domain expression of equation (1) by Fast Fourier Transform is:
Xi(k,l)＝Si(k,l)+Di(k,l)+Dw,i(k,l),i＝1,2 (3)
wherein, Xi(k,l)、Si(k,l)、Di(k, l) and Dw,i(k, l) are each xi(n)、si(n)、di(n) and dw,i(n) th frame kth band short time spectrum.
In performing S3, the following parameters are calculated according to the short-time spectrum of each frequency band:
s301, calculating the cross-correlation function of the sound signals between the microphones according to the short-time spectrums of the frequency bands.
Specifically, the ear microphone M1 and the ear microphone are providedThe device M2 has a distance d according to the sound signal x picked up by the ear microphone M1 and the ear microphone M21(n) and x2(n) (or two paths of wave beam output signals) each frequency band short-time spectrum carries out time delay and DOA estimation, and a common method is a generalized correlation method[7]. The cross-correlation function of the sound signal between the two extra-aural microphones is:
wherein, tau is time delay, NFFT is FFT point number, fSFor sampling the frequency, k, of the microphone signalLAnd kHUpper and lower limits, W, of the band range are selected for calculation, respectively12(k, l) are weighting coefficients based on the spectral information, and subscripts 1 and 2 denote the serial numbers of the extra-aural microphones. Weighting factor W12The value of (k, l) can be obtained by a weighting method based on signal-to-noise ratio, and can also be obtained by a weighting method based on phase transformation.
S302, according to the cross-correlation function
An estimate of the angle of incidence of the sound source between the two extra-aural microphones is calculated.
In particular, at τ ∈ [ -d/c, d/c]To the cross correlation function in the range of
Calculating and obtaining the value of the maximum value of the cross-correlation function corresponding to the time delay:
according to the value of the maximum value of the cross-correlation function obtained by the formula (5) corresponding to the time delay, calculating the estimated value of the sound source incident angle corresponding to the time delay as
Where c is the speed of sound.
Because when the target sound source is positioned right in front of the microphone, the target sound sourceIncident angle theta s0, so if the sound source incident angle θ of the combination of the ear microphone M1 and the ear microphone M2 is equal to12(l) When the estimated angle is also around 0 degrees, the current speech segment can be considered as the speech segment of the wearer; when the other directions interfere, the composite sound source incidence angle theta12(l) Will be biased in the other direction. According to the combined sound source incident angle theta12(l) And the estimated value can judge whether the current voice segment is an environmental noise segment or a voice segment of a wearer.
And S303, calculating the coherence among the microphones according to the cross-correlation function.
In order to determine whether it is wind noise, it is necessary to calculate the absolute value of the coherence between the ear microphone M1 and the ear microphone M2, where the value of the coherence between the ear microphone M1 and the ear microphone M2 is:
where it is a small quantity greater than 0, the divide by zero operation is avoided. The subscripts i and j indicate the serial numbers of the extra-aural microphones, and in this embodiment, i is 1 or 2, and j is 1 or 2.
Self-power spectrum and cross-power spectrum: when i ═ j, Rij(k,l)＝Rii(k,l)＝Rjj(k, l) is the self-power spectrum; when i ≠ j, it is a cross-power spectrum. Alpha is a fixed smoothing factor, and in order to reduce estimation bias, the value of alpha should be close to 1, but this will cause distortion in the initial section of speech and audio, especially when the wind noise energy is high, it will cause distortion of speech and audio signals for a long time, so the value of alpha is generally in the range of 0 to 1, and the value of alpha is typically 0.8.
S304, the incidence angle variance of the sound source incidence angles of the two extra-aural sound transmitters is calculated. Specifically, multiple frequency bands are combined into one sub-band, for example, one sub-band is combined every 500Hz, sound source localization is performed on each sub-band, and the incident angle θ of the sound source of each sub-band is obtainedij(κ, l), wherein κ represents a subband pointer;
estimating theta from sub-band sound source incident anglesij(κ, l), calculating the incidence angle variance:
wherein the content of the first and second substances,
κupis the sub-band upper limit.
S305, calculating a relative transfer function of the in-ear feedback microphone and the out-of-ear microphone, wherein the calculation steps are as follows:
assuming that there is only one in-ear feedback microphone or other type of vibration sensor, the frequency domain expression of equation (2) is:
wherein, Xin(k,l)、Sin(k,l)、Din(k, l) and
are each xin(n)、sin(n)、din(n) and
the kth frequency band short time spectrum of the l frame.
Calculating the relative transfer function of the in-ear feedback microphone and the out-of-ear microphone according to the short-time spectrums of the frequency bands as follows:
wherein, RTFi(k, l) is the transfer function of the ith out-of-ear microphone to the in-ear feedback microphone. RTF in non-wind-noise environments when the extra-aural microphone is relatively close in distanceiThe modulus of (k, l) for different values of i is close to a constant. To reduce the estimated variance of the relative function, a relative transfer function is calculated that fuses the multiple frequency bandsThe modulus value of (d):
wherein the coefficient β typically takes a value of 0.5, 1 or 2.
In the embodiment of the application, for the helmet wearing state detection under the condition of only having the binaural external sound transmitters, when S3-S4 are executed, the coherence among the binaural external sound transmitters is calculated according to the short-time spectrums of the frequency bands, and whether the absolute value of the coherence meets the numerical condition that the current frame is wind noise is judged; synthesizing at least 2 frequency band short-time spectrums into a sub-frequency band short-time spectrum; estimating the incidence angle of the sound source for each sub-band short-time spectrum, and calculating the variance of the incidence angle; judging whether the incidence angle variance meets the numerical condition that the current frame is wind noise; and when both meet the numerical condition of wind noise, determining the sound signal as wind noise. In order to reduce misjudgment, whether the current frame is the voice of the wearer can be judged according to the composite estimated value of the incident angle of the sound source, and if the current frame is not the voice of the wearer, the judgment of wind noise or environmental noise is carried out.
As shown in fig. 8, the following steps S801 to S805 are specifically performed:
s801, calculating and obtaining a composite sound source incidence angle estimated value theta of sound signals picked up by the ear-to-ear sound transmitter M1 and the ear-to-ear sound transmitter M2 in a certain frequency band range according to the formula (4) and the formula (5)12(l) Determining the composite sound source incident angle theta12(l) Whether greater than the angle threshold thetaThIf the combined sound source angle of incidence θ12(l) Greater than an angle threshold thetaThThe current frame is not the voice of the wearer, and the angle threshold thetaThCan take the value of 20 degrees;
s802, according to the formula (6) and the formula (7), judging whether the current frame is wind noise, and judging the coherence | C between the ear-external microphone M1 and the ear-external microphone M2 when the current frame is the wind noiseijWhen the difference value of the value of (k, l) | and the value of 0 is smaller than the coherence threshold and the value of the incidence angle variance VAR (theta (l)) is larger than the variance threshold, the wind noise is obtained; when the microphone is coherent | CijThe difference between the value of (k, l) | and the value of 1 is less than the coherence threshold, and the variance of the angle of incidenceWhen VAR (theta (l)) is smaller than the variance threshold, wind noise is not generated;
s803, when the current frame is judged not to be the voice of the wearer and to be wind noise through S801 and S802, the wind noise state value VAD of the current frame is markedwind(l) 1 is ═ 1; otherwise, marking the wind noise state value VAD of the current framewind(l)＝0；
S804, counting the wind noise state value VAD in a period of timewind(l) Frame number of 1, i.e.:
equation (11) can also be implemented by means of first-order smoothing, that is:
NumOfFrm(l)＝ηNumOfFrm(l-1)+(1-η)VADwind(l) (12)
the value of eta is close to 1 as much as possible, and the calculation amount can be effectively reduced by adopting the formula (12) and the storage space is reduced at the same time.
S805, when the number of frames NumOfFrm (l) with the wind noise state value of 1 in a period of time exceeds the set frame number threshold, the helmet is not worn in the traveling state currently; otherwise, the helmet is worn in a traveling state.
In the embodiment of the application, for the traveling state helmet wearing state detection under the condition of one out-of-ear microphone and one in-ear feedback microphone, when S3-S4 is executed, the coherence between the out-of-ear microphone and the in-ear feedback microphone is calculated according to the short-time spectrums of the frequency bands, and whether the absolute value of the coherence between the microphones meets the numerical condition that the current frame is wind noise is judged; calculating a module value of a relative transfer function fused with at least 2 frequency band short-time spectrums, and judging whether the module value is in accordance with the numerical condition of wind noise; and when both are in accordance with the numerical condition of wind noise, determining the current frame of the sound signal as the wind noise.
Specifically, as shown in fig. 9, the following steps S901 to S905 are performed:
s901, calculating the absolute value | C of the coherence of the out-of-ear microphone Mi and the in-ear feedback microphone Mj according to the formula (6)ij(k, l) |, when | CijWhen (k, l) | is close to 1, i.e. | CijIf the difference between (k, l) and 1 is less than the coherence threshold, then it is not wind noise at this moment, and the value VAD of the first wind noise statewind,1(l)＝0；|CijWhen (k, l) | is close to 0, i.e. | CijThe difference between (k, l) and 0 is less than the coherence threshold, which is the wind noise, and the value VAD of the first wind noise statewind,1(l)＝1；
S902, calculating a relative transfer function module value | RTF (kappa, l) | of the out-of-ear microphone Mi and the in-ear feedback microphone Mj according to a formula (10), wherein when the module value | RTF (kappa, l) | exceeds a certain threshold, the typical value of the threshold is 10 to 100, wind noise is generated, and the VAD value of the second wind noise state is obtained at the momentwind,2(l) 1 is ═ 1; when the module value | RTF (κ, l) | is smaller, the value VAD for the second wind noise statewind,2(l)＝0；
S903, VAD according to the first wind noise statewind,1(l) And second wind noise State VADwind,2(l) The value of (A) determines whether the current frame is wind noise, and in order to reduce false detection by using AND operation, namely the wind noise state VAD of the current framewind(l)＝VADwind,1(l)&VADwind,2(l) When the values of the first wind noise state and the second wind noise state are both 1, the wind noise state VAD of the current framewind(l) Set to 1, otherwise set to 0.
S904, counting the wind noise state VAD in a period of time according to the formula (11) or the formula (12)wind(l) The frame number numoffrm (l) of 1;
s905, VAD value of wind noise state in a period of timewind(l) When the number of frames NumOfFrm (l) which is 1 exceeds the set frame number threshold, the helmet is not worn in the traveling state currently; otherwise, the helmet is worn in a traveling state.
In the embodiment of the application, for the traveling state helmet wearing state detection under the condition that a double-ear external microphone and an in-ear feedback microphone exist, any one of the above schemes can be adopted, the two schemes can be fused, and the detection performance can be generally improved through the fusion. As shown in fig. 10, the specific fusion method is as follows:
s1001, adopting the advancing state helmet wearing state detection steps S801-S804 under the condition of the binaural external sound transmitter to count a first wind noise frame number which is recorded as NumOfFrm1(l)；
S1002, adopting a detection step of the traveling state of an ear-external microphone and an ear-internal feedback microphone and detecting the helmet wearing state S901-S904 to count the number of second wind noise frames which are recorded as NumOfFrm2(l)；
S1003, combining the first wind noise frame NumOfFrm1(l) And a second wind noise frame NumOfFrm2(l) The two can be added to obtain the total noise frame number:
NumOfFrm(l)＝NumOfFrm1(l)+NumOfFrm2(l) (13)
s1004, judging whether NumOfFrm (l) exceeds a threshold, if so, judging that the helmet is not worn in the traveling state, otherwise, judging that the helmet is worn in the traveling state.
In a non-traveling state, the helmet has a function of isolating external noise, so that the signal-to-noise ratio is high and the wind noise is weak. In addition, the action of the helmet can alter the transfer function of the wearer to the extra-aural microphone.
In the embodiment of the application, for the wearing state detection of the non-advancing helmet only with the binaural external microphone, the speech endpoint detection is performed based on the binaural external microphone, and a method based on the signal-to-noise ratio or a deep learning network method based on a small model can be adopted. The method for detecting the non-traveling state helmet wearing state of the binaural external sound transmitters comprises the steps of calculating the absolute value of coherence among the binaural external sound transmitters according to the short-time spectrums of the frequency bands; accumulating absolute values of coherence of the microphones over a full frequency band of the sound signal; and judging whether the helmet is worn in the current non-travelling state or not according to the accumulation result.
As shown in fig. 11 in particular, the following steps S1101-S1103 are performed:
s1101, calculating an absolute value | C of coherence of the intra-aural microphone M1 and the extra-aural microphone M2 of the current frame according to the formula (6)ij(k,l)|；
S1102, calculating | C in the current frame frequency band rangeij(k, l) | accumulated value, i.e.:
wherein k islowAnd kupA lower band limit and an upper band limit, respectively, a general lower band limit klowCorresponding to a frequency of 125Hz, an upper band limit of kupCorresponding to a frequency of 3750 Hz;
s1103, judging whether the accumulated value C (l) is larger than a certain threshold CthIf the judgment result is yes, the helmet is worn in the non-travelling state currently, otherwise the helmet is not worn in the non-travelling state currently;
in order to improve the detection accuracy, whether the accumulated values C (l) of the multiple frames are all larger than C or not can be judgedthAnd counting is carried out, and a total judgment is carried out according to the result of multiple frames.
In an embodiment of the application, a method for non-traveling state helmet fit detection with only one out-of-ear microphone and one in-ear feedback microphone includes calculating a modulus of a relative transfer function of the out-of-ear microphone to the in-ear feedback microphone from the respective frequency band short-time spectra; obtaining typical values under the condition of not wearing the helmet, wherein the first typical value is a relative transfer function module value of environmental noise, and the second typical value is a relative transfer function module value of the voice of a wearer; if the current frame is the voice of the wearer, judging whether the module value of the relative transfer function and the second typical value meet the numerical condition of wearing the helmet or not, and judging whether the helmet is worn in the current non-advancing state or not; and if the current frame is the environmental noise, judging whether the module value of the relative transfer function and the first typical value meet the numerical condition of wearing the helmet or not, and judging whether the helmet is worn in the current non-advancing state or not.
As shown in fig. 12, the following steps S1201-S1204 are performed:
s1201, calculating a module value | RTF (kappa, l) | of a relative transfer function from the ear external microphone to the ear internal feedback microphone according to a formula (10);
s1202, typical values of the absolute value of RTF (kappa, l) of the helmet are obtained, wherein the first typical value is a relative transfer function module value absolute value of the environmental noise RTFn(κ, l) |, second exemplaryRelative transfer function module value RTF with value of wearer's voices(κ,l)|；
S1203, judging whether the current frame is the voice of the wearer; if it is not the voice of the wearer, the relative transfer function value | RTF (k, l) | is compared with the first typical value | RTFn(κ, l) | are compared; if it is the wearer's voice, the band relative transfer function value | RTF (k, l) | and the second typical value | RTFs(κ, l) | are compared; specifically, an in-ear feedback microphone may be utilized for voice endpoint detection;
s1204, if the | RTF (k, l) | is equal to the first typical value | RTFnThe difference between (k, l) | is larger than the set threshold, or the band relative transfer function value | RTF (k, l) | and the second typical value | RTFsIf the difference value of (kappa, l) is larger than a set threshold value, the helmet is worn in a non-travelling state at present; otherwise, the helmet is not worn in the non-travelling state.
For the detection of the wearing state of the helmet in a non-traveling state with the double-ear external microphone and the single-ear internal microphone, any one of the two methods can be adopted, the two methods can be fused, and the detection performance can be improved through the fusion.
Regarding the judgment of whether the helmet is in a traveling state when being worn, an adopted method comprises the step of determining whether the helmet is in the traveling state when being worn and detected according to GPS positioning information by combining other sensors such as a GPS positioner on the helmet or a GPS positioning system of a smart phone. Or the wind noise of the sound signals picked up by the left and right ear microphones can be directly adopted for joint judgment. When the traveling state judgment is performed using the sound signal picked up by the microphone, the criterion is as follows: if the wind noise characteristics of the left ear and the right ear are close to the same, the helmet is in a traveling state during wearing detection; since the wind direction is irregular, if the wind noise characteristics of the left and right ears are significantly different, the helmet is in a non-traveling state when worn for detection.
The helmet wearing automatic detection method provided by the embodiment of the application is based on the shielding effect of the helmet on wind noise, the isolation effect on external environment noise and the influence on the voice of a wearer, helmet wearing detection is carried out through coherent calculation and relative transfer function calculation, and experimental results show that the detection performance is extremely high, the detection result is accurate, and application requirements are met.
It will be further appreciated by those of ordinary skill in the art that the elements and algorithm steps of the examples described in connection with the embodiments disclosed herein may be embodied in electronic hardware, computer software, or combinations of both, and that the components and steps of the examples have been described in a functional general in the foregoing description for the purpose of illustrating clearly the interchangeability of hardware and software. Whether these functions are performed in hardware or software depends on the particular application of the solution and design constraints. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied in hardware, a software module executed by a processor, or a combination of the two. A software module may reside in Random Access Memory (RAM), memory, Read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
The above-mentioned embodiments, objects, technical solutions and advantages of the present application are described in further detail, it should be understood that the above-mentioned embodiments are merely exemplary embodiments of the present application, and are not intended to limit the scope of the present application, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present application should be included in the scope of the present application.
Claims (14)
1. A method for automatically detecting the wearing state of a helmet is characterized by comprising the following steps:
picking up sound signals of at least 2 microphones;
respectively carrying out fast Fourier transform on the sound signals of the at least 2 microphones to obtain short-time spectrums of each frequency band;
determining whether the current frame of the sound signal is wind noise according to the short-time spectrums of the frequency bands;
and calculating the wind noise frame number in the sound signal, and judging the wearing condition of the helmet in the advancing state according to the wind noise frame number.
2. The method of claim 1, further comprising:
calculating a cross-correlation function of sound signals among the microphones in the at least 2 microphones according to the short-time spectrums of the frequency bands;
calculating a composite sound source incident angle estimation value among all microphones in the at least 2 microphones according to the cross-correlation function;
and judging whether the current frame is the voice of the wearer according to the composite sound source incident angle estimated value, and judging wind noise or environmental noise if the current frame is not the voice of the wearer.
3. The method of claim 1 or 2, wherein the at least 2 microphones are binaural microphones; the determining whether the current frame of the sound signal is wind noise according to the short-time spectrums of the frequency bands comprises:
calculating coherence between the double-ear external sound transmitters according to the short-time spectrums of the frequency bands, and judging whether the absolute value of the coherence meets the numerical condition that the current frame of the sound signal is the wind noise;
synthesizing at least 2 frequency band short-time spectrums into a sub-frequency band short-time spectrum;
estimating the incidence angle of the sound source for each sub-band short-time spectrum, and calculating the variance of the incidence angle; judging whether the incidence angle variance meets the numerical condition that the current frame of the sound signal is wind noise;
and when both are in accordance with the numerical condition of wind noise, determining that the current frame of the sound signal is wind noise.
4. The method of claim 1, wherein the at least 2 microphones are an out-of-ear microphone and an in-ear feedback microphone; the determining whether the current frame of the sound signal is wind noise according to the short-time spectrums of the frequency bands comprises:
calculating coherence between an ear external microphone and an ear internal feedback microphone according to the short-time spectrums of the frequency bands, and judging whether the absolute value of the coherence between the microphones meets the numerical condition that the current frame of the sound signal is wind noise;
calculating a module value of a relative transfer function fusing at least 2 frequency band short-time spectrums, and judging whether the module value meets the numerical condition that the current frame of the sound signal is wind noise;
and when both are in accordance with the numerical condition of wind noise, determining that the current frame of the sound signal is wind noise.
5. The method according to claim 3 or 4, wherein the calculating of the wind noise frame number in the sound signal and the determining of the wearing condition of the helmet in the traveling state according to the wind noise frame number comprise:
counting the number of wind noise frames in the sound signal within a period of time;
judging whether the wind noise frame number meets the numerical condition that the helmet is not worn in the advancing state; if yes, the helmet is not worn in the traveling state at present.
6. The method of claim 1 or 2, wherein the at least 2 microphones are a binaural external microphone and an in-ear feedback microphone; the method further comprises the following steps:
calculating coherence between the double-ear external sound transmitters according to the short-time spectrums of the frequency bands, and judging whether the absolute value of the coherence meets the numerical condition that the current frame of the sound signal is the wind noise;
synthesizing at least 2 frequency band short-time spectrums into a sub-frequency band short-time spectrum;
estimating the incidence angle of the sound source for each sub-band short-time spectrum, and calculating the variance of the incidence angle; judging whether the incidence angle variance meets the numerical condition that the current frame of the sound signal is wind noise;
when both are in accordance with the numerical condition of wind noise, determining that the current frame of the sound signal is wind noise;
counting the number of wind noise frames of the sound signal within a period of time under the condition of the double-ear external sound transmitter to obtain a first number of wind noise frames;
calculating coherence between an out-of-ear microphone and an in-ear feedback microphone according to the short-time spectrums of the frequency bands, and judging whether the absolute value of the coherence between the microphones meets the numerical condition that the current frame of the sound signal is wind noise;
calculating a module value of a relative transfer function fusing at least 2 frequency band short-time spectrum frequency bands, and judging whether the module value meets the numerical condition that the current frame of the sound signal is wind noise;
when both of the two values meet the numerical condition that the current frame of the sound signal is wind noise, determining the current frame of the sound signal to be wind noise; counting the number of wind noise frames of the sound signal within a period of time under the conditions of an out-of-ear microphone and an in-ear feedback microphone to obtain a second number of wind noise frames;
and adding the values of the first wind noise frame number and the second wind noise frame number, and judging whether the helmet is worn in the current advancing state according to the addition result.
7. The method of claim 1, wherein the at least 2 microphones are binaural external microphones, the method further comprising the step of detecting a helmet fit in a non-travel state:
calculating the absolute value of the coherence between the double-ear external sound transmitters according to the short-time spectrums of the frequency bands;
accumulating absolute values of coherence of the microphones over a full frequency band of the sound signal;
and judging whether the accumulated result meets the numerical condition of wearing the helmet or not, and judging whether the helmet is worn in the current non-advancing state or not.
8. The method of claim 2, wherein the at least 2 microphones are an out-of-ear microphone and an in-ear feedback microphone, the method further comprising the step of detecting a helmet fit in a non-travel state:
calculating the module value of the relative transfer function from the ear external microphone to the ear internal feedback microphone according to the short-time spectrums of the frequency bands;
obtaining typical values under the condition of not wearing the helmet, wherein the first typical value is a relative transfer function module value of environmental noise, and the second typical value is a relative transfer function module value of the voice of a wearer;
if the current frame is the voice of the wearer, judging whether the module value of the relative transfer function and the second typical value meet the numerical condition of wearing the helmet or not, and judging whether the helmet is worn in the current non-advancing state or not;
and if the current frame is the environmental noise, judging whether the module value of the relative transfer function and the first typical value meet the numerical condition of wearing the helmet or not, and judging whether the helmet is worn in the current non-advancing state or not.
9. The method of claim 1, further comprising determining from the GPS signal whether a helmet fit detection is in a travel state.
10. The device for automatically detecting the wearing state of the helmet is characterized by comprising a microphone and a digital signal processing module on an earphone;
the microphone is used for picking up sound signals; the number of the microphones is at least 2;
the digital signal processing module is used for respectively carrying out fast Fourier transform on the sound signals of the at least 2 microphones to obtain short-time spectrums of each frequency band; determining whether the current frame of the sound signal is wind noise according to the short-time spectrums of the frequency bands; and calculating the number of wind noise frames in the sound signal, and judging the wearing condition of the helmet in the advancing state.
11. The apparatus of claim 10, wherein the microphones comprise an out-of-ear microphone and an in-ear feedback microphone.
12. The apparatus of claim 10, wherein the dsp module is further configured to calculate an absolute value of coherence between binaural external speakers according to the short-time spectra of the respective frequency bands, and determine whether a numerical condition for wearing a helmet in a non-traveling state is met.
13. The apparatus of claim 10, wherein the digital signal processing module is further configured to calculate a module value of a relative transfer function from the ear-external microphone to the ear-internal feedback microphone according to the short-time spectrum of each frequency band, and determine whether a numerical condition for wearing the helmet in a non-traveling state is satisfied.
14. The device of claim 10 or 11, further comprising a GPS locator for determining whether the helmet is worn in a travelling state.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010671306.5A CN111833890B (en) | 2020-07-13 | 2020-07-13 | Device and method for automatically detecting wearing state of helmet |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010671306.5A CN111833890B (en) | 2020-07-13 | 2020-07-13 | Device and method for automatically detecting wearing state of helmet |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111833890A true CN111833890A (en) | 2020-10-27 |
| CN111833890B CN111833890B (en) | 2023-07-25 |
Family
ID=72922831
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010671306.5A Active CN111833890B (en) | 2020-07-13 | 2020-07-13 | Device and method for automatically detecting wearing state of helmet |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111833890B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150058002A1 (en) * | 2012-05-03 | 2015-02-26 | Telefonaktiebolaget L M Ericsson (Publ) | Detecting Wind Noise In An Audio Signal |
| CN109215677A (en) * | 2018-08-16 | 2019-01-15 | 北京声加科技有限公司 | A kind of wind suitable for voice and audio is made an uproar detection and suppressing method and device |
| US20190064750A1 (en) * | 2015-09-04 | 2019-02-28 | 3M Innovative Properties Company | Personal protective equipment and methods of monitoring time of usage of personal protective equipment |
| CN110830870A (en) * | 2019-11-26 | 2020-02-21 | 北京声加科技有限公司 | Earphone wearer voice activity detection system based on microphone technology |
-
2020
- 2020-07-13 CN CN202010671306.5A patent/CN111833890B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150058002A1 (en) * | 2012-05-03 | 2015-02-26 | Telefonaktiebolaget L M Ericsson (Publ) | Detecting Wind Noise In An Audio Signal |
| US20190064750A1 (en) * | 2015-09-04 | 2019-02-28 | 3M Innovative Properties Company | Personal protective equipment and methods of monitoring time of usage of personal protective equipment |
| CN109215677A (en) * | 2018-08-16 | 2019-01-15 | 北京声加科技有限公司 | A kind of wind suitable for voice and audio is made an uproar detection and suppressing method and device |
| CN110830870A (en) * | 2019-11-26 | 2020-02-21 | 北京声加科技有限公司 | Earphone wearer voice activity detection system based on microphone technology |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111833890B (en) | 2023-07-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| TWI763727B (en) | Automatic noise cancellation using multiple microphones | |
| US10535362B2 (en) | Speech enhancement for an electronic device | |
| US9025782B2 (en) | Systems, methods, apparatus, and computer-readable media for multi-microphone location-selective processing | |
| JP6150988B2 (en) | Audio device including means for denoising audio signals by fractional delay filtering, especially for "hands free" telephone systems | |
| KR101275442B1 (en) | Systems, methods, apparatus, and computer-readable media for phase-based processing of multichannel signal | |
| US9338547B2 (en) | Method for denoising an acoustic signal for a multi-microphone audio device operating in a noisy environment | |
| EP2679022B1 (en) | Systems, methods, apparatus, and computer-readable media for spatially selective audio augmentation | |
| US11115762B2 (en) | Hearing device for own voice detection and method of operating a hearing device | |
| US8909523B2 (en) | Method and acoustic signal processing system for interference and noise suppression in binaural microphone configurations | |
| US20100046770A1 (en) | Systems, methods, and apparatus for detection of uncorrelated component | |
| WO2014051969A1 (en) | System and method of detecting a user's voice activity using an accelerometer | |
| JP2005522078A (en) | Microphone and vocal activity detection (VAD) configuration for use with communication systems | |
| CN109196877A (en) | On the head of personal voice equipment/head detection outside | |
| US11553286B2 (en) | Wearable hearing assist device with artifact remediation | |
| EP3869821B1 (en) | Signal processing method and device for earphone, and earphone | |
| CN103428609A (en) | Apparatus and method for removing noise | |
| EP4118648A1 (en) | Audio processing using distributed machine learning model | |
| KR20210109526A (en) | Specific sound detector and method, and program | |
| EP3799032B1 (en) | Audio system and signal processing method for an ear mountable playback device | |
| CN111833890B (en) | Device and method for automatically detecting wearing state of helmet | |
| US9277316B2 (en) | Sound processing device and sound processing method | |
| US20230197050A1 (en) | Wind noise suppression system | |
| EP4061019A1 (en) | A headset capable of compensating for wind noise | |
| US11743661B2 (en) | Hearing aid configured to select a reference microphone | |
| EP4287657A1 (en) | Hearing device with own-voice detection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |