A kind of method for recognizing sound-groove and device and without key bicycle lock system and implementation methodTechnical field
The present invention relates to field of speech recognition, particularly relate to a kind of method for recognizing sound-groove and device.
Background technology
Following two kinds are had in the utilization on the key of vehicle-mounted nothing:
(1) adopt RFID technique, keyless systems is not having to need in self-locking scope to be undertaken opening the door and closing the door by the ACK button on car door.
(2) adopt the keyless access system of fingerprint recognition, only can open the door with use fingerprint limit, a RFID device need not be carried again.
But, existing RFID technique without key, need to be undertaken opening the door and closing the door by the ACK button on car door within 3 meters, manual contact check pushbutton switch car door.Fingerprint recognition needs contact could control car door without key, and on hand because of being stained with glue etc. because of carelessness when None-identified fingerprint, just cannot ought open the door.In addition, these two kinds of modes all need manual palpation contact car successfully to open the door, and cannot realize every empty close door.
Summary of the invention
For solving problems of the prior art, the invention provides a kind of method of Application on Voiceprint Recognition.
The present invention includes following steps: S1: voice are collected; S2: voice pre-service; S3: vocal print characteristics extraction; S4: model comparision, by the vocal print eigenwert of speaker with compare stored in the vocal print eigenwert in pattern base in advance, if identical, execution voice command; If not identical, return and perform step S1.
The present invention is further improved, and in step S3, described vocal print eigenwert comprises acoustic feature, lexical characteristics, prosodic features, languages, dialect, accent, channel information.
The present invention is further improved, in step S4, described model comparision method is: first carry out identity according to voice kind, dialect and accent information and tentatively differentiate, then gets rid of recording by channel information and imitates, finally by acoustic feature, lexical characteristics and prosodic features determination speaker ' s identity.
Present invention also offers a kind of device realizing said method, comprising: voice collection module, for collecting voice; Voice pretreatment module, for carrying out pre-service to the voice collected so that extract vocal print eigenwert; Vocal print characteristics extraction module, for extracting the vocal print eigenwert in voice; Model comparision module, for by the vocal print eigenwert of speaker with compare stored in the vocal print eigenwert in pattern base in advance, if identical, execution voice command; If not identical, return and perform voice collection module.
The present invention is further improved, and the vocal print eigenwert that described vocal print characteristics extraction module is extracted comprises acoustic feature, lexical characteristics, prosodic features, languages, dialect, accent, channel information.
The present invention is further improved, in model comparision module, described model comparision method is: first carry out identity according to voice kind, dialect and accent information and tentatively differentiate, then get rid of recording by channel information and imitate, finally by acoustic feature, lexical characteristics and prosodic features determination speaker ' s identity.
Present invention also offers one without key bicycle lock system, comprise speech control module, without key bicycle lock device, also comprise voice print identification device, for passing through speech-input instructions switch lock.
The present invention is further improved, and describedly also comprises fingerprint identification device without key bicycle lock system.
Present invention also offers a kind of above-mentioned implementation method without key bicycle lock system, comprise the following steps: A1: by the voice messaging of car owner and driving encrypted message stored in pattern base; A2: voice are collected; A3: voice pre-service; A4: vocal print characteristics extraction; A5: model comparision, by the vocal print eigenwert of speaker with compare stored in the car owner's vocal print eigenwert in pattern base in advance, if identical, execution voice command; If not identical, return and perform steps A 2.
The present invention is further improved, in steps A 5, before voice command performs, perform a model comparision, then model comparision times N adds 1, if N is greater than 3, then speech control module prompting car owner phonetic entry password, judge that whether password is identical with the password in described pattern base, if so, then can direct door opening; If not, then execution steps A 2 is returned.
Compared with prior art, the invention has the beneficial effects as follows: speech recognition, without key, confirm switch gate every blank operation, convenient and swift, confirm identity mode by Application on Voiceprint Recognition simultaneously, guarantee to only have car owner just can carry out switch gate, be conducive to automotive safety.
Accompanying drawing explanation
Fig. 1 is method for recognizing sound-groove process flow diagram of the present invention;
Fig. 2 is the vehicle-mounted keyless systems structural representation of the present invention;
Fig. 3 is the present invention's vehicle-mounted keyless systems implementation method process flow diagram.
Embodiment
Below in conjunction with drawings and Examples, the present invention is described in further details.
As shown in Figure 1, method for recognizing sound-groove of the present invention comprises the following steps: S1: voice are collected; S2: voice pre-service; S3: vocal print characteristics extraction; S4: model comparision, by the vocal print eigenwert of speaker with compare stored in the vocal print eigenwert in pattern base in advance, if identical, execution voice command; If not identical, return and perform step S1.
In step S1, it is realized by microphone that voice are collected, and in step S3, described vocal print eigenwert comprises acoustic feature, lexical characteristics, prosodic features, languages, dialect, accent, channel information.In step S4, described model comparision method is: first carry out identity according to voice kind, dialect and accent information and tentatively differentiate, then gets rid of recording by channel information and imitates, finally by acoustic feature, lexical characteristics and prosodic features determination speaker ' s identity.
Correspond, voice print identification device of the present invention comprises: voice collection module, for collecting voice; Voice pretreatment module, for carrying out pre-service to the voice collected so that extract vocal print eigenwert; Vocal print characteristics extraction module, for extracting the vocal print eigenwert in voice; Model comparision module, for by the vocal print eigenwert of speaker with compare stored in the vocal print eigenwert in pattern base in advance, if identical, execution voice command; If not identical, return and perform voice collection module.
Voice collection module is microphone.The vocal print eigenwert that described vocal print characteristics extraction module is extracted comprises acoustic feature, lexical characteristics, prosodic features, languages, dialect, accent, channel information.In model comparision module, described model comparision method is: first carry out identity according to voice kind, dialect and accent information and tentatively differentiate, then get rid of recording by channel information and imitate, finally by acoustic feature, lexical characteristics and prosodic features determination speaker ' s identity.
Of the present invention is increase speech control module, voice print identification device and microphone at existing automotive interior without on key bicycle lock device without key bicycle lock system, for by speech-input instructions switch lock.Microphone is responsible for collecting voice, voice print identification device is responsible for speech analysis process, Application on Voiceprint Recognition etc., if determine voice recognition commands, such as we customize enabling and two voice commands of closing the door, not having in self-locking scope without key bicycle lock system, when identifying these two orders, confirm that this is the sound that car owner sends simultaneously, switch gate is confirmed that order is issued without key bicycle lock system by this, during without key bicycle lock system acceptance to switch gate order, then perform corresponding action, for car owner opens or closed door.
In addition, can arrange fingerprint identification device of the present invention without in key bicycle lock system, automobile also can be supported, and fingerprint recognition is opened the door, how a kind of approach of door opening.
In implementation method of the present invention, before use speech recognition equipment, to need the voice messaging of car owner and driving encrypted message, stored in pattern base, so that system only recognizes the voice of car owner, to guarantee to only have car owner just can carry out switch gate, be conducive to automotive safety.
As the present invention's another embodiment without the implementation method of key bicycle lock system, in model comparision step, before voice command performs, pattern base stores the number of times that execution pattern compares, and performs a model comparision, then model comparision times N adds 1, if N is greater than 3, then speech control module prompting car owner phonetic entry password, judges that whether password is identical with the password in described pattern base, if so, then can direct door opening; If not, then return execution voice and collect step.Increase password driving lock-step rapid, car owner's situation that voice print identification device is not recognized because of flu or other reasons aphonia can be avoided.
The embodiment of the above is better embodiment of the present invention; not limit specific embodiment of the invention scope with this; scope of the present invention comprises and is not limited to this embodiment, and all equivalence changes done according to the present invention are all in protection scope of the present invention.