CN104900228B - A kind of recognition methods of suspicious enabling sound - Google Patents
A kind of recognition methods of suspicious enabling sound Download PDFInfo
- Publication number
- CN104900228B CN104900228B CN201510218727.1A CN201510218727A CN104900228B CN 104900228 B CN104900228 B CN 104900228B CN 201510218727 A CN201510218727 A CN 201510218727A CN 104900228 B CN104900228 B CN 104900228B
- Authority
- CN
- China
- Prior art keywords
- sound
- module
- suspicious
- enabling
- acquisition
- 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.)
- Active
Links
Abstract
The invention discloses a kind of identification device of suspicious enabling sound and recognition methods, identification device includes sequentially connected voice triggering module, sound acquisition module, sound characteristic extraction process module, sound recognition module and alarm module.Sound collection is carried out when being greater than threshold value, the sound of acquisition is pre-processed and is converted into digital signal;Feature extraction is carried out to digital signal and obtains characteristic parameter, obtains feature vector F to be identified after eliminating magnitude differencesi';Calculate Fi' feature vector F ' corresponding with the enabling sound prestoredI is depositedSimilarity Si, according to SiThe classification of sample to be identified is determined, therefore, it is determined that whether the sound of acquisition is suspicious enabling sound.The present invention, to reinforce the antitheft of door lock, is thus improved the safety of door lock, can effectively take precautions against that there may be the illegal enabling behaviors dramatically different with normal door sound by the identification of suspicious enabling sound.
Description
Technical field
The present invention relates to voice recognition technologies, and in particular to a kind of identification device for being exclusively used in suspicious enabling sound and identification
Method belongs to door lock technical field of burglary prevention to improve the tamper-security of door lock.
Background technique
As home intelligence is gradually popularized, intelligent refrigerator, Intelligent cleaning robot, smart television etc. have stepped into usually old
Common people family, but intelligentized security door does not enter market also.With people household and other property safeties are realized it is continuous
Enhancing, to antitheft, the anti-snatch and anti-access control system pounded intelligence more stringent requirements are proposed.
Currently, most of household antitheft doors, car door etc. are all made of key unlocking, therefore a current design direction is exactly
The structure of lock is improved to further increase its safety.Although the design optimization of lock can effectively reduce some unlockings and steal
Event, but still have the case for pilferage of largely unlocking or force open the door.Its reason is:(1) improvement of above-mentioned lock is still kept away
The risk that unavoidable thief is opened the door using unlocking tool or skeleton key;(2) it can not detect that other abnormal door open events occur,
It switchs, for example forces open the door as destructive;Based on this, the normal and suspicious sound for opening the door and (referring to the sound such as theft enabling) is known
It is not the critical issue and technology of intelligent anti-theft door, is with a wide range of applications and practical value.
Other than common door lock, the door identified using speech recognition technology to password or speaker's identity is had also appeared
Taboo technology, the technology is although highly-safe, but equally exists some shortcomings:(1) most users get used to traditional lock
It opens the door with key, is not also to be good at receiving speech recognition technology, it is wideless to popularize face.(2) system complex, price is high, main at present
Based on important unit use, the less use of ordinary residence;(3) influence of the speech recognition technology vulnerable to speaker's state,
Recognition accuracy is difficult to ensure.
Summary of the invention
In view of the above shortcomings of the prior art, the object of the present invention is to provide a kind of identification of suspicious enabling sound dresses
It sets and recognition methods, the present invention, to reinforce the antitheft of door lock, is to existing door lock safety by the identification of suspicious enabling sound
Useful supplement, can be further improved the safety of door lock.Existing various door locks can be used, adaptable.
Purpose to realize the present invention uses following technical scheme:
A kind of identification device of suspicious enabling sound, including sequentially connected voice triggering module, sound acquisition module, sound
Sound feature extraction processing module, sound recognition module and alarm module;
Voice triggering module:The size for detecting ambient sound, starts sound collection when sound energy value is greater than the set value
Module;
Sound acquisition module:Ambient sound data are acquired, and carry out analog-to-digital conversion;
Sound characteristic extraction process module:Various processing are carried out to mention to the voice data that sound acquisition module sends over
Take related sound feature;
Sound recognition module:By what is stored in sound characteristic and database that sound characteristic extraction process module sends over
The corresponding sound characteristic of normal door sound is compared, and calculates its similarity, to determine whether for suspicious sound;
Alarm module:If the sound that sound recognition module confirmly detects is suspicious sound, alarm module issues scene report
Alert, onsite alarming is that buzzer warning or/and alarm lamp are alarmed.
It further, further include photo module and transmission information module;The output of sound recognition module is agreed to play according to mould simultaneously
Block and transmission information module;
Photo module:If the sound that sound recognition module confirmly detects is suspicious sound, start photo module to outdoors
Carry out video capture;
Send information module:If the sound that sound recognition module confirmly detects is suspicious sound, sending information module will
Relevant information is sent to specified communication by way of wireless transmission and receives terminal, and the relevant information is the photo captured
Or suspicious enabling short message prompt;It is mobile phone or/and computer that communication, which receives terminal,.
A kind of suspicious enabling sound identification method, identification step is as follows,
S1, ambient sound energy size is detected every setting time, sound collection is carried out when greater than threshold value;
S2, the voice data of acquisition is passed through into band-pass filter, subsequently into A/D conversion module, by voice data
Digital signal is converted by analog signal;
S3, characteristic parameter is obtained to the digital signal progress feature extraction that step S2 is obtained, then combines characteristic parameter
To the feature vector for constituting higher-dimension together, and the calculating of following formula is carried out to each dimensional feature vector, eliminate magnitude differences:
Wherein, FiFor i-th of the original feature vector actually obtained, FimaxAnd FiminRespectively represent database prestores i-th
The maximum value and minimum value of a feature vector, Fi' it is i-th of feature vector to be identified eliminated after magnitude differences;
S4, the feature vector F to be identified after eliminating magnitude differences is calculatedi' corresponding with the enabling sound that database prestores
Feature vector F 'I is depositedSimilarity Si, to all SiAscending sort is carried out, preceding K minimum S is utilizediSample class carry out
The ballot that the minority is subordinate to the majority, K are odd number, determine the classification of sample to be identified, therefore, it is determined that whether the sound of acquisition is suspicious
Enabling sound;
S5, when for suspicious enabling sound, then issue corresponding alarm.
In step S5, when for suspicious enabling sound, while photo module is triggered to taking pictures outdoors, and will be taken pictures
Piece is sent to specified communication in a manner of being wirelessly transferred and receives terminal, and it is mobile phone or/and computer that communication, which receives terminal,.
The feature wherein extracted in step S3 includes:Enabling duration T;Sound number of segment N;The time zone that sound occurs
Between;Sound amplitude maximum value, minimum value, mean value, variance, energy histogram;The frequency domain character of sound, frequency domain character is using quickly
Fourier transformation obtains.
Wherein similarity calculation is carried out using Euclidean distance formula in step S4, and calculation formula is as follows:
Si=| | Fi'-F′I is deposited||
Or
(σ is preset normal parameter).
Compared with the conventional method, the present invention has the advantages that:
1) present invention by the identification of suspicious enabling sound to reinforce the antitheft of door lock, if there is suspicious enabling sound,
Then thus automatic alarm or transmission prompt information improve the safety of door lock to related personnel.
2) present invention can effectively take precautions against there may be the illegal enabling behavior dramatically different with normal door sound, as by force
Row forces open the door, improper key unlocking etc., and these illegal enabling behaviors are also difficult to successfully manage at present.
3) present invention may apply to various door locks, adaptable, are the useful supplements to existing door lock safety.
Detailed description of the invention
Fig. 1-identification device structural block diagram of the present invention.
Fig. 2-identification process figure of the present invention.
Specific embodiment
The present invention carries out automatic detection using mode identification method, and identification device is adopted by voice triggering module 110, sound
Collect module 120, sound characteristic extraction process module 130, sound recognition module 140, alarm module 150, photo module 160, hair
Information module etc. is sent to constitute, as shown in Figure 1.
Voice triggering module:The size for detecting ambient sound, starts acquisition module when sound energy value is greater than the set value;
Sound acquisition module:Voice data is acquired, and carries out analog-to-digital conversion;Voice triggering module and sound acquisition module are used
Arriving is same set of sound collection element, sound is acquired first with acquisition module, by collected voice input to sounds trigger
Module is judged, if it is greater than the initiation value of setting, is fed back to sound acquisition module and is continued to acquire current speech data and mould
Otherwise number conversion is removed the data that front acquires, is acquired to sound next time.
Sound characteristic extraction process module:Pretreatment and shear treatment are carried out to voice data, and extract related sound spy
Sign;
Sound recognition module:By what is stored in sound characteristic and database that sound characteristic extraction process module sends over
The corresponding sound characteristic of normal door sound is compared, and calculates its similarity, to determine whether for suspicious sound;
Alarm module:If identification module is determined as suspicious sound, which will start buzzer and alarm lamp;If it is not
Method molecule is opening the door, which effectively will prevent it from continuing malfeasance;
Photo module:While starting alert program, start photo module to video capture is carried out outdoors, to leave door
Image at that time outside, the information of people especially in image;Photo module is typically mounted on the surface of door, can to certain area into
Row is taken pictures;
Send information module:The module sends the photo captured to the mobile phone of nominator by way of wireless transmission
Or on computer, it is allowed to know there are suspicious enabling situation at the first time, to judge whether strictly illegally to open the door by photo, with
Decide whether to take measures.Simultaneously in the case where confirmation is illegal opens the door, which provides help for the processing of subsequent case.
To prevent criminal from destroying to the identification device, in practical applications, sound acquisition module can be installed
Near the lock core and inside built-in introduction chamber, be both convenient for the abundant acquisition of dual lock process sound, at the same also effectively avoid by
Malicious sabotage.Voice triggering module, alarm module are mountable indoors Anywhere, note abnormalities convenient for indoor member.Sound
Feature extraction processing module, sound recognition module, send information module mainly realized by software design, therefore, Ke Yi
The integrated application software of personal-machine interactive mode is developed on computer, while can also be developed mobile phone app and be communicated with computer.
Photo module is that criminal is avoided to find, pin hole camera shooting can be installed on door, controls the integrated software (or app) by developing
It completes and shows.
The recognition methods of the suspicious enabling sound of the present invention, specifically includes following steps, referring also to Fig. 1 and Fig. 2:,
S1, voice triggering module are greater than every the energy size of time interval (such as 0.1 second) the detection ambient sound of setting
Start sound acquisition module when threshold value;
Energy balane formula is:
Wherein, s (n) represents the voice signal of n-th of sampled point, and L represents the points that time interval is sampled.Sampling number
Equal to sample frequency (generally higher than 4KHz) multiplied by time interval (such as 0.1 second).
S2, sound acquisition module by speech simulation signal by band-pass filter, subsequently into A/D conversion module,
It is converted into digital signal;
S3, feature extraction is carried out to digital signal using sound characteristic extraction process module, the feature of extraction mainly includes:
Enabling duration T, sound number of segment N, the time interval that sound occurs, sound amplitude maximum value, minimum value, mean value, variance, energy
Histogram, the frequency domain character of sound are measured, frequency domain character is obtained using Fast Fourier Transform (FFT);Then characteristic parameter is combined to one
The feature vector for constituting higher-dimension is acted, and carries out the calculating of following formula to each dimensional feature, eliminates magnitude differences:
Wherein, the F of right side of the equal signiFor i-th of the original feature vector actually obtained, FimaxAnd FiminRespectively represent data
The maximum value and minimum value for the ith feature vector that library prestores, the F of left side of the equal signi' for after i-th elimination magnitude differences
Feature vector to be identified.
S4, feature vector F to be identified is calculated using sound recognition modulei' the enabling sound that prestores with database is being (comprising just
Hold-open door sound and improper enabling sound) corresponding each feature vector F 'I is depositedSimilarity Si, wherein F 'I is depositedAlso according to step
Rapid S3 has carried out the processing of elimination magnitude differences;Similarity can be used Euclidean distance formula and be calculated, and calculation formula is as follows:
Si=| | Fi'-F′I is deposited||
Or
(σ is preset normal parameter).
To all SiAscending sort is carried out, preceding odd number (such as 3 or 5) minimum S is utilizediSample class carry out
The ballot that the minority is subordinate to the majority determines the classification of sample to be identified.
S5, when the classification of ballot is suspicious, then trigger photo module to taking pictures outdoors, and start warning device,
Such as buzzer will finally take a picture and be sent to preset mobile phone terminal by wireless transmission method (such as wifi),
Decide whether to alarm by owner.
Finally, it is stated that above-described embodiment is only used to illustrate the technical scheme of the present invention rather than limits, although referring to compared with
Good embodiment describes the invention in detail, those skilled in the art should understand that, it can be to skill of the invention
Art scheme is modified or replaced equivalently, and without departing from the objective and range of technical solution of the present invention, should all be covered at this
In the scope of the claims of invention.
Claims (2)
1. a kind of theft preventing method of door lock, it is characterised in that:It is realized by a kind of suspicious enabling voice recognition device, it is described suspicious
Enabling voice recognition device include sequentially connected voice triggering module, sound acquisition module, sound characteristic extraction process module,
Sound recognition module and alarm module;
Voice triggering module:The size for detecting ambient sound, starts sound acquisition module when sound energy value is greater than the set value;
Sound acquisition module:Ambient sound data are acquired, and carry out analog-to-digital conversion;
Sound characteristic extraction process module:The voice data that sound acquisition module sends over is carried out at pretreatment and shearing
Reason, to extract related sound feature;
Sound recognition module:It is normal by being stored in sound characteristic and database that sound characteristic extraction process module sends over
The corresponding sound characteristic of enabling sound is compared, and calculates its similarity, to determine whether for suspicious sound;
Alarm module:If the sound that sound recognition module confirmly detects is suspicious sound, alarm module issues onsite alarming, existing
Field alarm is that buzzer warning or/and alarm lamp are alarmed;
Specific identification step is as follows,
S1, ambient sound energy size is detected every setting time, sound collection is carried out when greater than threshold value;
S2, the voice data of acquisition is passed through into band-pass filter, subsequently into A/D conversion module, by voice data by mould
Quasi- signal is converted into digital signal;
S3, characteristic parameter is obtained to the digital signal progress feature extraction that step S2 is obtained, characteristic parameter is then combined to one
The feature vector for constituting higher-dimension is acted, and carries out the calculating of following formula to each dimensional feature vector, eliminates magnitude differences:
Wherein, FiFor i-th of the original feature vector actually obtained, FimaxAnd FiminRespectively represent i-th of spy that database prestores
Levy the maximum value and minimum value of vector, Fi' it is i-th of feature vector to be identified eliminated after magnitude differences;
The feature extracted in step S3 includes:Enabling duration T;Sound number of segment N;The time interval that sound occurs;Sound width
Spend maximum value, minimum value, mean value, variance, energy histogram;The frequency domain character of sound, frequency domain character are become using fast Fourier
Change acquisition;
S4, the feature vector F to be identified after eliminating magnitude differences is calculatedi' corresponding with the enabling sound that database prestores feature
Vector Fi'It depositsSimilarity Si, to all SiAscending sort is carried out, preceding K minimum S is utilizediSample class carry out a small number of clothes
From most ballots, K is odd number, determines the classification of sample to be identified, therefore, it is determined that whether the sound of acquisition is suspicious enabling sound
Sound;
Similarity calculation is carried out using Euclidean distance formula, and calculation formula is as follows:
Si=| | Fi'-Fi'It deposits||
Or
(σ is preset normal parameter);
S5, when for suspicious enabling sound, then issue corresponding alarm.
2. a kind of theft preventing method of door lock according to claim 1, it is characterised in that:The suspicious enabling voice recognition dress
Set further includes photo module and transmission information module;The output of sound recognition module agrees to play lighting module simultaneously and sends information mould
Block;
Photo module:If the sound that sound recognition module confirmly detects is suspicious sound, starting photo module to carrying out outdoors
Video capture;
Send information module:If the sound that sound recognition module confirmly detects is suspicious sound, sending information module will be related
Information is sent to specified communication by way of wireless transmission and receives terminal, the relevant information be the photo captured or
Person/and suspicious enabling short message prompt;
In step s 5, when for suspicious enabling sound, while photo module is triggered to taking pictures outdoors, and will be taken a picture
It is sent to specified communication in a manner of wireless transmission and receives terminal, it is mobile phone or/and computer that communication, which receives terminal,.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510218727.1A CN104900228B (en) | 2015-04-30 | 2015-04-30 | A kind of recognition methods of suspicious enabling sound |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510218727.1A CN104900228B (en) | 2015-04-30 | 2015-04-30 | A kind of recognition methods of suspicious enabling sound |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104900228A CN104900228A (en) | 2015-09-09 |
CN104900228B true CN104900228B (en) | 2018-11-16 |
Family
ID=54032859
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510218727.1A Active CN104900228B (en) | 2015-04-30 | 2015-04-30 | A kind of recognition methods of suspicious enabling sound |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104900228B (en) |
Families Citing this family (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105336330A (en) * | 2015-10-15 | 2016-02-17 | 上海易景信息科技有限公司 | Simple voice recognizing system |
CN106296890B (en) * | 2016-07-22 | 2019-06-04 | 北京小米移动软件有限公司 | Unlocking method, device and the mobile terminal of mobile terminal |
CN107730829A (en) * | 2016-10-20 | 2018-02-23 | 北京奥斯达兴业科技有限公司 | Alarm interlock method and device |
CN107833322A (en) * | 2017-11-07 | 2018-03-23 | 惠州市瑞时智控科技有限公司 | A kind of Intelligent door control system and its control method based on audio identification |
CN112950876A (en) * | 2018-04-28 | 2021-06-11 | 侯苏华 | Anti-theft method, anti-theft device, anti-theft terminal and computer readable medium |
CN108877814B (en) * | 2018-05-23 | 2020-12-29 | 中南林业科技大学 | Inspection well cover theft and damage detection method, intelligent terminal and computer readable storage medium |
CN110969805A (en) * | 2018-09-30 | 2020-04-07 | 杭州海康威视数字技术股份有限公司 | Safety detection method, device and system |
CN109472311A (en) * | 2018-11-13 | 2019-03-15 | 北京物灵智能科技有限公司 | A kind of user behavior recognition method and device |
CN110047509B (en) * | 2019-03-28 | 2021-04-02 | 国家计算机网络与信息安全管理中心 | Two-stage subspace partitioning method and device |
CN110148423A (en) * | 2019-04-18 | 2019-08-20 | 广东思派康电子科技有限公司 | A kind of computer readable storage medium and antitheft interactive voice speaker can be used for |
CN111188536A (en) * | 2020-02-13 | 2020-05-22 | 德施曼机电(中国)有限公司 | Anti-theft system and method for intelligent lock |
CN112489330B (en) * | 2020-12-02 | 2022-02-11 | 广东电网有限责任公司江门供电局 | Warehouse anti-theft alarm method |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1889172A (en) * | 2005-06-28 | 2007-01-03 | 松下电器产业株式会社 | Sound sorting system and method capable of increasing and correcting sound class |
EP2192562A1 (en) * | 2008-11-26 | 2010-06-02 | ASSA ABLOY Sicherheitstechnik GmbH | Emergency door control device and door system |
CN103021043A (en) * | 2011-09-22 | 2013-04-03 | 邱金和 | Monitoring system |
CN103927808A (en) * | 2014-04-30 | 2014-07-16 | 中国矿业大学(北京) | Voiceprint-recognition-based system and method for detecting uniqueness of coal mine entering worker |
-
2015
- 2015-04-30 CN CN201510218727.1A patent/CN104900228B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN104900228A (en) | 2015-09-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104900228B (en) | A kind of recognition methods of suspicious enabling sound | |
US6744462B2 (en) | Apparatus and methods for resolution of entry/exit conflicts for security monitoring systems | |
CN206515931U (en) | A kind of face identification system | |
CN105469484B (en) | A kind of APP smart homes lock | |
CN105279872A (en) | Anti-theft monitoring device and monitoring system for entrance door | |
CN104881911A (en) | System And Method Having Biometric Identification Instrusion And Access Control | |
KR101962753B1 (en) | Digital Door System for preventing Crime and Robbery Preventing Method by Using thereof | |
CN101839092A (en) | Anti-theft alarming lock capable of shooting pictures and sending information | |
CN107633580A (en) | A kind of smart lock and its control method, intelligent coffer | |
CN205080681U (en) | Register one's residence door electronic monitoring device | |
WO2019228091A1 (en) | Attendance management method and attendance management device | |
CN106781141A (en) | A kind of mall entrance intelligent safety and defence system | |
CN112991585A (en) | Personnel entering and exiting management method and computer readable storage medium | |
CN109922206A (en) | It is a kind of for the intelligent alarm method, apparatus of mobile phone and including its system | |
CN104809780A (en) | Access control method based on fingerprint recognition and face recognition | |
CN104835222A (en) | Entrance guard control method based on voice recognition, facial recognition and fingerprint recognition | |
CN110580762A (en) | Intelligent household access control system and method | |
CN108166850A (en) | Intelligent alarm theftproof door lock system | |
CN103375106A (en) | Antitheft door structure and method capable of realizing intelligent antitheft judgment effect by aid of computer | |
CN104809784A (en) | Access control method based on voice recognition and face recognition | |
CN209912036U (en) | Face identification security entrance guard all-in-one | |
CN106530563A (en) | Antitheft system of computer information | |
CN110853192A (en) | Unlocking and locking control method of safe intelligent lock | |
CN109360363A (en) | A kind of Intelligent monitoring anti-theft system and method | |
CN203849779U (en) | Face identification idling door lock structure |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |