Summary of the invention
The voice acquisition method, device, one kind that the purpose of the application is to provide a kind of intelligent wearable device are computer-readableStorage medium and a kind of intelligent wearable device can start phonetic incepting group under the premise of actively not touching intelligent wearable devicePart improves the convenience of data under voice process.
In order to solve the above technical problems, the application provides a kind of voice acquisition method of intelligent wearable device, which is adoptedSet method includes:
When receiving bone vibration data, digital signal corresponding with bone vibration data is generated；
Judge whether the amplitude of digital signal is greater than preset value；
If so, starting phonetic incepting component, and voice data is acquired using phonetic incepting component.
Optionally, before whether the amplitude for judging digital signal is greater than preset value, further includes:
The noise information of current environment is obtained, and preset value is determined according to noise information；Wherein, the vibration width of noise informationValue is positively correlated with preset value.
Optionally, further includes:
The remaining capacity of intelligent wearable device is obtained, and judges whether remaining capacity is less than default electricity；
If so, adjusting preset value according to remaining capacity；Wherein, the remaining capacity of intelligent wearable device and preset value negativeIt closes.
Optionally, after using phonetic incepting component acquisition voice data, further includes:
When the corresponding digital signal amplitude of current bone vibration data is less than preset value, current bone vibration data is recordedCorresponding digital signal amplitude continuously less than or equal to preset value duration；
Judge whether duration is greater than preset duration；
If so, closing phonetic incepting component.
Optionally, further includes:
When starting phonetic incepting component, control audio player module plays the first prompt tone；
When closing phonetic incepting component, control audio player module plays the second prompt tone.
Optionally, after using phonetic incepting component acquisition voice data, further includes:
Voice data is sent to speech recognition module, so that speech recognition module executes speech recognition behaviour to voice dataMake.
Optionally, when starting phonetic incepting component, further includes:
Record the bone vibration data in target time section；Wherein, target time section is the Startup time of phonetic incepting componentTo the close moment corresponding period；
Bone vibration data in target time section is sent to speech recognition module with voice data, so as to speech recognition mouldRoot tuber is according to the bone vibration data and voice data execution speech recognition operation in target time section.
Optionally, further includes:
When receiving the recognition failures information of speech recognition module transmission, control audio player module plays voice numberAccording to so that wearer checks voice data.
Optionally, when the bone vibration data in target time section is sent to speech recognition module with voice data, alsoInclude:
Calculate the delay duration of the Startup time of phonetic incepting component and the initial time of voice data；
Delay duration is sent to speech recognition module, so that speech recognition module is right before executing speech recognition operationBone vibration data in target time section executes intercept operation and obtains the bone vibration number for having identical initial time with voice dataAccording to.
Present invention also provides a kind of voice acquisition device of intelligent wearable device, which includes:
Analog-to-digital conversion module, for generating digital signal corresponding with bone vibration data when receiving bone vibration data；
Amplitude judgment module, for judging whether the amplitude of digital signal is greater than preset value；
Voice acquisition module, for starting phonetic incepting component, and utilize when the amplitude of digital signal is greater than preset valuePhonetic incepting component acquires voice data.
Present invention also provides a kind of computer readable storage mediums, are stored thereon with computer program, computer programThe step of voice acquisition method of above-mentioned intelligent wearable device executes is realized when execution.
Present invention also provides a kind of intelligent wearable device, including memory and processor, calculating is stored in memoryMachine program, processor realize that the voice acquisition method of above-mentioned intelligent wearable device executes when calling the computer program in memoryThe step of.
This application provides a kind of voice acquisition methods of intelligent wearable device, including when receiving bone vibration data,Generate digital signal corresponding with bone vibration data；Judge whether the amplitude of digital signal is greater than preset value；If so, starting languageSound receiving unit, and voice data is acquired using phonetic incepting component.
The application generates the corresponding digital signal of bone vibration data when receiving bone vibration data, passes through judgement number letterNumber amplitude be greater than preset value and judge whether bone vibration data meets preset standard.Since wearer is bound to companion when speakingWith the vibration for having bone, therefore can be using bone vibration signal as judging wearer whether in the reference conditions spoken.Due to wearingThe noise of person's local environment can equally make the bone of wearer there are certain vibration, and the application is made only by the way that preset value is arrangedHave the amplitude for receiving bone vibration data and the corresponding digital signal of bone vibration data be greater than preset value when just start voice connectComponent is received, realizes the purpose using phonetic incepting component acquisition voice data.During the data under voice of the application, nothingWearer's hand touches device is needed, as long as bone vibration data when wearer speaks and speaks meets relevant criterion and can startPhonetic incepting component.Therefore, the application can start phonetic incepting component under the premise of actively not touching intelligent wearable device,Improve the convenience of data under voice process.The application additionally provides a kind of voice collecting dress of intelligent wearable device simultaneouslyIt sets, a kind of computer readable storage medium and a kind of intelligent wearable device, there is above-mentioned beneficial effect, details are not described herein.
Specific embodiment
To keep the purposes, technical schemes and advantages of the embodiment of the present application clearer, below in conjunction with the embodiment of the present applicationIn attached drawing, the technical scheme in the embodiment of the application is clearly and completely described, it is clear that described embodiment isSome embodiments of the present application, instead of all the embodiments.Based on the embodiment in the application, those of ordinary skill in the artEvery other embodiment obtained without making creative work, shall fall in the protection scope of this application.
Current intelligent wearable device all has the function of voice collecting, the voice that intelligent wearable device passes through acquisition wearerData realize the multiple functions such as call, speech recognition, vocal print unlock.Current intelligent wearable device generally requires wearer's touchingOr double-click intelligent wearable device and just can star phonetic incepting component, but the above method needs wearer to contact intelligent wearing groupPart then cannot achieve voice collecting when wearer is inconvenient to contact intelligent wearable device.There is also intelligence to wear in the prior artThe voice collecting scheme that equipment is always maintained at phonetic incepting component open state is worn, but always on phonetic incepting component is to functionThe requirement of consumption is very high, and intelligent wearable device can be allowed to consume whole electricity in a short time.Therefore, this mode is although protect alwaysIntelligent wearable device can not be contacted voice collecting can be realized by holding the unlatching of phonetic incepting component, but will will lead to equipment continuation of the journeyAbility wretched insufficiency.Various deficiencies based on the above-mentioned prior art during intelligent wearable device acquires voice data, hereinThe voice acquisition method of new intelligent wearable device is provided by following embodiment, improves data under voice stream to reachThe purpose of the convenience of journey.
Below referring to Figure 1, Fig. 1 is a kind of voice acquisition method of intelligent wearable device provided by the embodiment of the present applicationFlow chart.
Specific steps may include:
S101: when receiving bone vibration data, digital signal corresponding with bone vibration data is generated；
Wherein, the executing subject of the present embodiment can be the processing chip of intelligent wearable device, do not limit intelligence herein and wearThe type of equipment is worn, intelligent wearable device can be intelligent glasses, head-mounted display, wireless Bluetooth headsets etc..
Intelligent wearable device application wearable technology carries out intelligentized design to daily wearing, develops and can dressThe general name of equipment, therefore intelligent wearable device can directly be contacted with wearer, intelligent wearable device can pass through various sensingsDevice acquires the physiologic information (such as heart rate information, body temperature information, bone vibration information) or intelligent wearable device local environment of wearerEnvironmental information (such as noise information, temperature and humidity information).
The bone vibration data being previously mentioned in this step can be bone conduction transducer data collected, therefore receive bone vibrationBefore dynamic data, there may be the operations of acquisition bone vibration data for default, that is to say, that exists and utilizes bone conduction transducer (boneSensor the process of bone vibration data) is acquired.The intelligent wearable device that the present embodiment is mentioned can be include bone conduction transducerIntelligent wearable device.
It should be noted that the bone vibration data that the present embodiment is mentioned is the position that wearer wears intelligent wearable deviceBone vibration data does not limit which physical feeling that bone vibration data derives from wearer herein.As feasible embodiment,Since bone vibration data is the reference data for judging whether to start phonetic incepting component, bone vibration data can be wearerThe data that bone vibrates more apparent physical feeling are generated when speaking, and are such as set close to skull or upper mandibular equipotential.
It is understood that cause intelligent wearable device wearer bone generate vibration factor can be divided into it is inherent becauseElement and external factor, internal factor refer to the factor that the certain physiological activities of wearer itself cause bone to vibrate, such as: playing sprayIt sneezes, cough, speak；External factor refers to the factor for vibrating bone in wearer's local environment, such as: outer playback sound,The roar etc. of car engine.The origin for generating bone vibration data is not limited in S101, as long as intelligent wearable device can be madeWearing at bone vibration.
After receiving bone vibration data, it is corresponding bone vibration signal can be generated by way of ADC analog-to-digital conversionDigital signal, so as to carry out it is subsequent analysis come judge bone vibration data whether meet open phonetic incepting component related itemPart.Specifically, there may be bone conduction transducers to acquire bone vibration data simultaneously in the intelligent wearable device that the present embodiment is previously mentionedThe bone vibration data of acquisition is sent to the operation of processing chip, then the phase in S101, S102 and S103 is executed by processing chipClose operation.Bone conduction transducer is not limited herein for the acquisition mode of bone vibration data, bone conduction transducer can be according to default weekPhase stage acquires bone vibration data, can also acquire bone vibration data with duration.Certainly, in order to realize the real-time of voice collectingProperty, bone conduction transducer can acquire bone vibration data in real time and the bone vibration data acquired in real time is sent to processing chip.
S102: judge whether the amplitude of digital signal is greater than preset value；If so, into S103；If it is not, then terminating to flowJourney；
Wherein, the purpose of this step is to judge whether the corresponding digital signal of bone vibration signal meets preset condition, such asFruit, which meets preset condition then, can carry out starting phonetic incepting component described in S103 and acquire the operation of voice data, such asFruit, which does not meet preset condition, can then terminate process flow described in the present embodiment.As a kind of feasible embodiment, whenIt, can be after the process for terminating the present embodiment again through bone conduction when this step judges the amplitude of digital signal no more than preset valueSensor acquires new bone vibration data and is judged next time.
It is understood that bone conduction transducer can collect bone vibration when the wearer of intelligent wearable device speaksData, when wearer's bone caused by the noise in environment vibrates, bone conduction transducer can equally collect bone vibration data.But in identical environment wearer speak with it is silent when collected bone vibration data be different, specifically, identicalThe bone vibration signal generated when wearer is silent under environment is only environmental noise effect as a result, generating when wearer speaksBone vibration signal is the stack result that wearer's physiological vibrations add environmental noise effect.Therefore, the generation when wearer speaksThe corresponding digital signal amplitude of bone vibration data be greater than only environmental noise generate the corresponding data-signal of bone vibration dataAmplitude.The preset value that this step is previously mentioned is for distinguishing the bone vibration data generated when wearer speaks and only ambient noiseWhen the amplitude of bone vibration data that generates, therefore preset value can flexibly be set according to the specifically used environment of intelligent wearable deviceIt sets, the specific value of preset value is not defined herein.When the amplitude of digital signal is greater than preset value, determine that wearer works asThe preceding moment is speaking and can star phonetic incepting component to acquire voice data.When the amplitude of data-signal is less than or equal toWhen the preset value, it is possible to determine that wearer does not speak at current time, without opening phonetic incepting component.
The specific value for not limiting the preset value in the present embodiment, it is default there may be being set or adjusted before S102The operation of value, if the preset value in this step be can distinguish bone vibration data when intelligent wearable device wearer speaks andThe value of bone vibration data when wearer does not speak.Further, preset value, which is equivalent to, judges what whether wearer spokeReference standard, preset value mean that more greatly this is more harsh with reference to judgment criteria, and the present embodiment can also be by adjusting preset valueSize adjustment judges the harsh degree whether wearer speaks.It will later about the detailed process for being arranged and adjust the preset valueEmbodiment be introduced.
It should be noted that the phonetic incepting component of the present embodiment default intelligent wearable device is not in normally opened mode etc.To the arriving of voice signal, phonetic incepting component is constantly in closed state before S103, only meets the judgement in S102Just start phonetic incepting component in the case where condition, can effectively save electric energy in this way, promotes the continuation of the journey of intelligent wearable deviceAbility.
S103: starting phonetic incepting component, and voice data is acquired using phonetic incepting component.
Wherein, this step is established in S102 it is determined that the amplitude of the corresponding digital signal of bone vibration data is greater than preset valueOn the basis of, start the phonetic incepting component that closed state is constantly in front of this step first.Default in the present embodimentThere are the interactions on signal for phonetic incepting component and the processing chip of intelligent wearable device, such as: processing chip is in judgement numberThe amplitude of signal is generated when being greater than preset value and is interrupted to start phonetic incepting component.Qualifier sound does not receive in the present embodimentThe specific type of component, as long as acquiring the device of voice data after capable of receiving the enabled instruction for handling chip can be seen asThe phonetic incepting component that the present embodiment is previously mentioned.
It is understood that there may be processing chip generating enabling signal between S102 and S103 and by enabling signalIt is sent to the operation of phonetic incepting component.After starting phonetic incepting component, phonetic incepting component can be acquired in certain timeVoice data, generate and shutdown signal and send shutdown signal there may also be processing chip after acquiring voice dataUnnecessary power consumption is reduced to the operation of phonetic incepting component to close phonetic incepting component in time.
Illustrate the treatment process of S101, S102 and S103 by the example in practical application: bone conduction transducer is experienced firstIt is vibrated to bone conduction, digital signal is converted to by ADC, which is conveyed to DA14195 chip (i.e. intelligent wearable deviceProcessing chip).When DA14195 chip judges the amplitude of digital information greater than preset value, DA14195 chip and mic (i.e. languageSound receiving unit) pin of connection is converted to high level by low level, generates to interrupt and inputs to mic, and mic is received on after interruptionEstablish beginning acquisition voice data by cable.
The present embodiment generates the corresponding digital signal of bone vibration data when receiving bone vibration data, passes through judgement numberThe amplitude of signal is greater than preset value to judge whether bone vibration data meets preset standard.Since wearer is bound to when speakingWith the vibration of bone, therefore can be using bone vibration signal as judging wearer whether in the reference conditions spoken.Due to wearingThe noise of wearer's local environment can equally make the bone of wearer there are certain vibration, and the present embodiment is made by the way that preset value is arrangedIt obtains and only just starts language when the amplitude for receiving bone vibration data and the corresponding digital signal of bone vibration data is greater than preset valueSound receiving unit realizes the purpose using phonetic incepting component acquisition voice data.In the data under voice mistake of the present embodimentCheng Zhong is not necessarily to wearer's hand touches device, as long as bone vibration data when wearer speaks and speaks meets relevant criterion meetingPhonetic incepting component can be started.Therefore, the present embodiment can start language under the premise of actively not touching intelligent wearable deviceSound receiving unit improves the convenience of data under voice process.
As the further supplement of embodiment corresponding for Fig. 1, there can be setting preset value before S102Operation: the noise information of current environment is obtained, and preset value is determined according to noise information；Wherein, the vibration amplitude of noise informationIt is positively correlated with preset value.
In order to improve the accuracy rate for judging whether wearer is speaking, the vibration of bone caused by needing to speak to wearer andThe vibration of bone caused by ambient noise distinguishes, and needs to set suitable number for preset value if wanting effectively distinguishValue.The operation of above-mentioned setting preset value gives the scheme of accurately setting preset value: first obtaining wearer and is presently in ringThe noise information in border determines preset value according to the power of ambient noise information.There are noise informations for default in the provisioning processThe corresponding relationship of vibration amplitude and preset value is operated according to the setting that the corresponding relationship can complete preset value, so as to work as front ringThe vibration amplitude and preset value of the noise information in border are positively correlated.
Of course, it is possible to which multiple amplitudes area will be divided between the maxima and minima of the vibration amplitude of ambient noise in advanceBetween, its bigger corresponding preset value of section intermediate value in the corresponding preset value in each amplitude section and amplitude section is also bigger.It canWith the corresponding amplitude section of noise information vibration amplitude of the determination current environment before S102, according to amplitude section determination pairThe preset value answered.
It is operated by the setting of above-mentioned preset value, can achieve following effect: when wearer is in quiet environment, in advanceIf value is relatively small, wearer, which whispers, can start the acquisition that phonetic incepting component completes voice data；At wearerWhen in noisy environment, preset value is relatively large, and wearer, which needs to speak up, just can star phonetic incepting component completion languageThe acquisition of sound data.Further, the vibration of bone caused by wearer's sound of speaking is bigger is more obvious, and is pacified if current environment is oppositeQuiet, wearer, which whispers, to be collected by phonetic incepting component, if current environment is relatively noisy, wearer needs to declaimWords could be collected by phonetic incepting component.Therefore, preset value is arranged according to the noisy degree of the noise information of current environment, noBut it can effectively distinguish bone caused by the vibration of bone caused by wearer speaks and ambient noise to vibrate, also help and connect according to voiceThe voice data for receiving component acquisition determines the speech content of wearer.
Specifically, the noise information of current environment can determine that detailed process is as follows by environmental noise learning process: firstLesser preset value is first set, starts phonetic incepting component when the amplitude of ambient noise is greater than the preset value, if receiveWhen effective voice messaging being not present in voice data, then it can be determined that the bone vibration signal being currently received is drawn by environmental noiseIt rises, preset value can be set as according to current environment noise information at this time, so that the amplitude when ambient noise is less than or equal to newlyPreset value.It is, of course, also possible to there is the operation for actively setting preset value, it may be assumed that when receiving preset value reset instruction,The noise information for obtaining current environment, it is corresponding with the correspondence setting noise information of preset value default according to noise informationValue.
The further supplement of above-mentioned embodiment corresponding for Fig. 1 is to wear from the angle of judgment accuracy to the intelligence of Fig. 1The improvement for wearing the voice acquisition method of equipment, refers to Fig. 2 below, and Fig. 2 is that a kind of adjustment provided by the embodiment of the present application is pre-If the flow chart of the method for value；The corresponding embodiment of Fig. 2 proposes preset value from the angle of intelligent wearable device cruising abilityAdjusted Option.The present embodiment is the method for adjustment to the preset value mentioned in the corresponding embodiment of Fig. 1, can by the present embodiment withThe corresponding embodiment of Fig. 1 combines to obtain more preferably embodiment, and specific steps may include:
S201: the remaining capacity of intelligent wearable device is obtained；
S202: judge whether remaining capacity is less than default electricity；If so, into S203；If it is not, then terminating process；
S203: preset value is adjusted according to remaining capacity；Wherein, the remaining capacity of intelligent wearable device and preset value negativeIt closes.
In general, intelligent wearable device is designed with power module, for providing for each device in intelligent wearable deviceElectric energy.The remaining capacity mentioned in the present embodiment refers specifically to the current residual electricity of the power module of intelligent wearable device, hereinThe type of power module is not limited, and power module can be rechargeable lithium battery, or solar panel.
Wherein, judge in S202 remaining capacity whether be less than default electricity meaning be determine intelligent wearable device electricityEnergy state, illustrates that intelligent wearable device is currently at low battery state when remaining capacity is less than default electricity, i.e. intelligence wearingThe current available duration of equipment is less than target duration.The specific value of default electricity is not defined in the present embodiment, as long asIllustrate that intelligent wearable device is in low battery state when remaining capacity is less than default electricity, as a kind of optional implementationScheme, default electricity can be 20 the percent of intelligent wearable device maximum residual electricity.The default electricity can also basisPractical application scene is adjusted flexibly, and higher default electricity such as is arranged in the working time of wearer, in wearer's sleeping timeLower default electricity is set.
It should be noted that under the premise of remaining capacity is less than default electricity, remaining capacity and preset value adjustedFor negative correlativing relation, i.e. lower bigger, the remaining electricity that needs to start the corresponding digital signal amplitude of phonetic incepting component of remaining capacityMeasure that higher to need to start the corresponding digital signal amplitude of phonetic incepting component smaller.Setting through this embodiment can reduce intelligenceBecause of the probability of ambient noise starting phonetic incepting component when energy wearable device remaining capacity is lower, needed when remaining capacity is lowerWearer just can star phonetic incepting component with bigger sound.Therefore when remaining capacity is lower using the present embodiment providesPreset value Adjusted Option can be improved the acquisition probability of efficient voice data (i.e. the voice data of wearer), reduce due to ringThe probability of border noise error starting can effectively extend the continuation of the journey duration of intelligent wearable device.
As the further supplement of embodiment corresponding for Fig. 1, voice data is being acquired using phonetic incepting componentLater, further include that following four phonetic incepting component closes mode:
(1) first kind phonetic incepting component closes mode:
Step 1: when the corresponding digital signal amplitude of current bone vibration data is less than preset value, recording current bone vibrationThe dynamic corresponding digital signal amplitude of data continuously less than or equal to preset value duration；
Step 2: judging whether duration is greater than preset duration；If so, closing phonetic incepting component.
There are certain time interval (usually 500ms), the first above-mentioned voices between two words when normal person speaksStandard time interval of the preset duration that receiving unit closing mode is mentioned between two words, when duration, which is greater than, to be presetWhen long, illustrate that wearer has piped down, phonetic incepting component can be closed.For example, if preset duration is 500ms, the2300ms detects that the corresponding digital signal amplitude of bone vibration data is less than preset value, then may determine that from 2300ms to theWhether the corresponding digital signal amplitude of the bone vibration data of any time is greater than preset value between 2800ms, if otherwise can closePhonetic incepting component.
(2) second class phonetic incepting components close mode:
Step 1: recording the duration of voice data, and judge whether the duration of voice data is greater than standard duration；If so,Enter step 2；If it is not, then entering step 1；
Step 2: judging whether the corresponding digital signal amplitude of current bone vibration data is less than standard value；If so, intoEnter step 3；If it is not, then entering step 2；
Step 3: closing phonetic incepting component.
Second class phonetic incepting component close the supplement of corresponding for Fig. 1 embodiment of mode by the duration of voice data andWhen the corresponding digital signal amplitude of prebone vibration data is as reference value, only it is greater than standard duration in voice data duration and works asPreceding bone vibration data corresponding digital signal amplitude just closes phonetic incepting component when being less than preset value.Above-mentioned closing voice connectsReceive the condition of component it is possible to prevente effectively from due to when wearer speaks two words bring there are time interval or speech content are too longData under voice lack problem.
(3) third class phonetic incepting component closes mode:
After starting phonetic incepting component, processing chip can recorde the starting duration of phonetic incepting component, work as startingShutdown signal is generated when duration is greater than preset duration and shutdown signal is sent to phonetic incepting component, so as to phonetic incepting componentThe voice data of preset duration can be acquired.
(4) the 4th class phonetic incepting components close mode:
After starting phonetic incepting component, the corresponding number of bone vibration signal that chip can be current with constant testing is handledWhether signal amplitude is greater than preset value, when detecting the corresponding digital signal amplitude of current bone vibration signal or be less than or equal toIllustrate that wearer has piped down when preset value, handling chip at this time can be generated shutdown signal and be sent to shutdown signalPhonetic incepting component, so as to phonetic incepting component can acquire wearer be in speak state when voice data.
It further, can be on the basis of above-mentioned supplement as advanced optimizing during voice collectingThere are following operations:
Operation 1: when starting phonetic incepting component, control audio player module plays the first prompt tone；
Operation 2: when closing phonetic incepting component, control audio player module plays the second prompt tone.
It is to play second to mention closing phonetic incepting component by playing the first prompt tone when starting phonetic incepting componentShow sound, helps wearer to understand the time point of voice collecting beginning and end, between Intelligent Optimal wearable device and wearerInteractivity promotes the user experience of wearer.
As the further supplement of embodiment corresponding for Fig. 1, voice data is being acquired using phonetic incepting componentLater, further include following operation: voice data is sent to speech recognition module, so that speech recognition module holds voice dataRow speech recognition operation.
Still there are numbers with intelligent wearable device for the device in intelligent wearable device for qualifier sound identification module hereinAccording to the device in interactive terminal device.When intelligent wearable device includes the speech recognition module, the place of intelligent wearable deviceManaging chip can be by way of wire transmission by the voice data transmission of acquisition to speech recognition module, by speech recognition moduleVoice data is identified to obtain recognition result, to execute the corresponding operation of recognition result.When the speech recognition module isWith intelligent wearable device there are when device in the terminal device of data interaction, the processing chip of intelligent wearable device can will be adoptedThe voice data of collection is transmitted to the terminal device including speech recognition module by way of wireless transmission, holds in terminal deviceRow speech recognition operation obtains recognition result, so that terminal device executes corresponding operation according to speech recognition result.Specifically,Speech recognition module after receiving voice data, can determine in phonetic reference model library with the matched key of the voice dataWord and using the keyword as recognition result.Further, the present embodiment compareing there may also be keyword and control commandRelationship generates control command according to the contrast relationship to execute relevant operation.
Fig. 3 is referred to below, and Fig. 3 is a kind of flow chart of audio recognition method provided by the embodiment of the present application；This realityApply example be to the processing operation executed in the corresponding embodiment of Fig. 1 when starting phonetic incepting component, can by the present embodiment withThe corresponding embodiment of Fig. 1 combines to obtain more preferably embodiment, and specific steps may include:
S301: the bone vibration data in target time section is recorded when starting phonetic incepting component；
Wherein, target time section is the Startup time of phonetic incepting component to the close moment corresponding period；
S302: the bone vibration data in target time section is sent to speech recognition module with voice data, so as to voiceIdentification module is according to the bone vibration data and voice data execution speech recognition operation in target time section.
Above-described embodiment by the object time bone vibration data and voice data be used as in speech recognition process toProcess object is sent to speech recognition module.Speech recognition module can be according to the bone vibration data and voice in target time sectionData generate keyword feature information；It can also determine the standard feature information pair in database with keyword feature information matchesThe keyword answered, to execute the corresponding operation of keyword.Bone vibration data and voice data is combined to carry out language in the present embodimentSound identifies the accuracy that can effectively improve speech recognition result, obtains relative to only speech recognition is carried out by voice dataRecognition result it is more reliable.
As advanced optimizing for Fig. 3 corresponding embodiment, can by target time section bone vibration data withVoice data executes following steps when being sent to speech recognition module:
The delay duration of the initial time of step 1, the Startup time for calculating phonetic incepting component and voice data；
Delay duration is sent to speech recognition module by step 2, so that speech recognition module is executing speech recognition operationIntercept operation is executed to the bone vibration data in target time section before and obtains the bone that there is identical initial time with voice dataVibration data.
Above-mentioned optimization is there are problems that delay phonetic incepting component acquisition voice data, because in processing chipDeterministic process needs certain time, then can also have delay to phonetic incepting component electrifying startup, and phonetic incepting component powers onArrive afterwards can steady operation equally exist delay.Therefore it should be noted that postponing bone vibration data caused by duration and voice data notThe problem of alignment, such as when wearer says " playing music ", bone vibration data is exactly complete Pinyin are as follows: bo fangYin yue, and voice data may are as follows: o fang yin yue needs to intercept data after buffered data, that is, needsGuarantee that two data are all o fang yin yue.
The treatment process that the above process is illustrated because of the speech detection algorithms in the DSP of processing chip needs 5.5ms,Then to phonetic incepting component electrifying startup there are the delay of 2.5ms, phonetic incepting component arrives after powering on is capable of steady operationDelay is 30ms, so phonetic incepting module data will about postpone bone vibration data 38ms.So in use, to infuseThe problem of meaning delay is misaligned, such as when wearer says that bone vibration data is exactly complete Pinyin are as follows: bo when playing musicFang yin yue, and voice data may be exactly that there are voice data are as follows: o fang yin yue is needed after buffered dataData are intercepted, that is, need to guarantee that two data are all o fang yin yue.The baud rate of phonetic incepting component is16Kbps, interception formula areLen be bone vibration data need to intercept by starting point of start time whenIt is long.
The difference of a length of bone vibration data and voice data initial time when delay mentioned above sends delay durationTo speech recognition module, so that speech recognition module intercepts bone vibration data according to delay duration, to identical initial timeBone vibration data and voice data carry out speech recognition operation.
As an alternative embodiment, carrying out speech recognition operation related data is sent to speech recognition moduleLater, there may also be following operations: when receiving the recognition failures information of speech recognition module transmission, control audio is playedComponent plays voice data, so that wearer checks voice data.
Defaulted herein when speech recognition module can not identify the data received or obtained recognition result is wrong, voiceIdentification module returns to recognition failures information to the processing chip of intelligent wearable device.When processing chip receives recognition failures informationThe voice data that control audio player module records before playing afterwards determines language so that wearer listens to speech content just nowThe reason of sound recognition failures, allows wearer to adjust the speech content of oneself to carry out the mistake of voice collecting and speech recognition againJourney.
Further, any one of the above embodiment description intelligent wearable device phonetic incepting component can forThe phonetic incepting component of decrease of noise functions can execute echo cancellation and/or noise reduction trim process to the voice data received.
For example, phonetic incepting component can built-in 2 microphones, at the bis- mic noise reduction algorithms progress noise reductions of alangoReason, a microphone, which is stablized, keeps clear call, another microphone physics actively eliminates noise, by collecting extraneous soundAfter being handled, the sound wave opposite with noise is issued, eliminates noise using principle of cancellation, is conducive to improve speech recognition belowDiscrimination.
Fig. 4 is referred to, Fig. 4 is a kind of knot of the voice acquisition device of intelligent wearable device provided by the embodiment of the present applicationStructure schematic diagram；
The apparatus may include:
Analog-to-digital conversion module 100, for when receiving bone vibration data, generating number letter corresponding with bone vibration dataNumber；
Amplitude judgment module 200, for judging whether the amplitude of digital signal is greater than preset value；
Voice acquisition module 300, for starting phonetic incepting component when the amplitude of digital signal is greater than preset value, andVoice data is acquired using phonetic incepting component.
The present embodiment generates the corresponding digital signal of bone vibration data when receiving bone vibration data, passes through judgement numberThe amplitude of signal is greater than preset value to judge whether bone vibration data meets preset standard.Since wearer is bound to when speakingWith the vibration of bone, therefore can be using bone vibration signal as judging wearer whether in the reference conditions spoken.Due to wearingThe noise of wearer's local environment can equally make the bone of wearer there are certain vibration, and the present embodiment is made by the way that preset value is arrangedIt obtains and only just starts language when the amplitude for receiving bone vibration data and the corresponding digital signal of bone vibration data is greater than preset valueSound receiving unit realizes the purpose using phonetic incepting component acquisition voice data.In the data under voice mistake of the present embodimentCheng Zhong is not necessarily to wearer's hand touches device, as long as bone vibration data when wearer speaks and speaks meets relevant criterion meetingPhonetic incepting component can be started.Therefore, the present embodiment can start language under the premise of actively not touching intelligent wearable deviceSound receiving unit improves the convenience of data under voice process.
Further, the voice acquisition device further include:
Preset value setup module determines preset value for obtaining the noise information of current environment, and according to noise information；ItsIn, the vibration amplitude and preset value of noise information are positively correlated.
Further, the voice acquisition device further include:
Preset value adjusts module, for obtaining the remaining capacity of intelligent wearable device, and judges whether remaining capacity is less thanDefault electricity；If so, adjusting preset value according to remaining capacity；Wherein, the remaining capacity Yu preset value of intelligent wearable device are negativeIt is related.
Further, the voice acquisition device further include:
Phonetic incepting component closedown module, for being preset when the current corresponding digital signal amplitude of bone vibration data is less thanWhen value, record the corresponding digital signal amplitude of current bone vibration data continuously less than or equal to preset value duration；AlsoFor judging whether duration is greater than preset duration；If so, closing phonetic incepting component.
Further, the voice acquisition device further include:
First cue module, for when starting phonetic incepting component, control audio player module to play the first prompt tone；
Second cue module, for when closing phonetic incepting component, control audio player module to play the second prompt tone.
Further, the voice acquisition device further include:
First data transmission blocks, for voice data to be sent to speech recognition module, so as to speech recognition module pairVoice data executes speech recognition operation.
Further, the voice acquisition device further include:
Bone vibration data logging modle, for recording the bone vibration data in target time section；Wherein, target time section isThe Startup time of phonetic incepting component is to the close moment corresponding period；
Second data transmission blocks are known for the bone vibration data in target time section to be sent to voice with voice dataOther module, so that speech recognition module executes speech recognition behaviour with voice data according to the bone vibration data in target time sectionMake.
Further, the voice acquisition device further include:
Module is checked, for controlling audio broadcasting group when receiving the recognition failures information of speech recognition module transmissionPart plays voice data, so that wearer checks voice data.
Further, the voice acquisition device further include:
Be delayed correction module, for calculating the delay of the Startup time of phonetic incepting component and the initial time of voice dataDuration；Be also used to postpone duration and be sent to speech recognition module, so as to speech recognition module execute speech recognition operation itThe preceding bone vibration data in target time section executes intercept operation and obtains shaking with the bone that voice data has identical initial timeDynamic data.
Since the embodiment of components of system as directed is corresponded to each other with the embodiment of method part, the embodiment of components of system as directed is askedReferring to the description of the embodiment of method part, wouldn't repeat here.
Present invention also provides a kind of computer readable storage mediums, have computer program thereon, the computer programIt is performed and step provided by above-described embodiment may be implemented.The storage medium may include: USB flash disk, mobile hard disk, read-only depositReservoir (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), magnetic disk orThe various media that can store program code such as CD.
Present invention also provides a kind of intelligent wearable devices, may include bone conduction transducer, phonetic incepting component, memoryAnd processor, bone conduction transducer is for receiving bone vibration data；Phonetic incepting component is for acquiring voice data；It is deposited in memoryThere is computer program that step provided by above-described embodiment may be implemented when processor calls the computer program in memory.Certain intelligent wearable device can also include various interfaces, power supply, audio player module, bluetooth module, speech recognition module etc.Component.
Each embodiment is described in a progressive manner in specification, the highlights of each of the examples are with other realitiesThe difference of example is applied, the same or similar parts in each embodiment may refer to each other.For system disclosed in embodimentSpeech, since it is corresponded to the methods disclosed in the examples, so being described relatively simple, related place is referring to method part illustration?.It should be pointed out that for those skilled in the art, under the premise of not departing from the application principle, alsoCan to the application, some improvement and modification can also be carried out, these improvement and modification also fall into the protection scope of the claim of this applicationIt is interior.
It should also be noted that, in the present specification, relational terms such as first and second and the like be used merely to byOne entity or operation are distinguished with another entity or operation, without necessarily requiring or implying these entities or operationBetween there are any actual relationship or orders.Moreover, the terms "include", "comprise" or its any other variant meaningCovering non-exclusive inclusion, so that the process, method, article or equipment for including a series of elements not only includes thatA little elements, but also including other elements that are not explicitly listed, or further include for this process, method, article orThe intrinsic element of equipment.Under the situation not limited more, the element limited by sentence "including a ..." is not arrangedExcept there is also other identical elements in the process, method, article or apparatus that includes the element.