CN109087655A - A kind of monitoring of traffic route sound and exceptional sound recognition system - Google Patents
A kind of monitoring of traffic route sound and exceptional sound recognition system Download PDFInfo
- Publication number
- CN109087655A CN109087655A CN201810851609.8A CN201810851609A CN109087655A CN 109087655 A CN109087655 A CN 109087655A CN 201810851609 A CN201810851609 A CN 201810851609A CN 109087655 A CN109087655 A CN 109087655A
- Authority
- CN
- China
- Prior art keywords
- sound
- frame
- data
- abnormal sound
- short
- 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.)
- Pending
Links
- 238000012544 monitoring process Methods 0.000 title claims abstract description 27
- 230000002159 abnormal effect Effects 0.000 claims abstract description 72
- 238000012545 processing Methods 0.000 claims abstract description 23
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 12
- 238000004891 communication Methods 0.000 claims abstract description 11
- 230000005236 sound signal Effects 0.000 claims abstract description 8
- 238000001228 spectrum Methods 0.000 claims description 18
- 238000000605 extraction Methods 0.000 claims description 17
- 238000005070 sampling Methods 0.000 claims description 13
- 238000009432 framing Methods 0.000 claims description 12
- 230000005540 biological transmission Effects 0.000 claims description 5
- 238000001514 detection method Methods 0.000 claims description 5
- 230000003542 behavioural effect Effects 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 230000003068 static effect Effects 0.000 claims description 3
- 230000005856 abnormality Effects 0.000 claims description 2
- 238000013528 artificial neural network Methods 0.000 abstract description 2
- 238000009434 installation Methods 0.000 abstract description 2
- 238000012549 training Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 10
- 206010039203 Road traffic accident Diseases 0.000 description 4
- 230000003321 amplification Effects 0.000 description 4
- 238000000034 method Methods 0.000 description 4
- 238000003199 nucleic acid amplification method Methods 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 2
- 230000001537 neural effect Effects 0.000 description 2
- 230000002547 anomalous effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
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
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/123—Traffic control systems for road vehicles indicating the position of vehicles, e.g. scheduled vehicles; Managing passenger vehicles circulating according to a fixed timetable, e.g. buses, trains, trams
- G08G1/127—Traffic control systems for road vehicles indicating the position of vehicles, e.g. scheduled vehicles; Managing passenger vehicles circulating according to a fixed timetable, e.g. buses, trains, trams to a central station ; Indicators in a central station
- G08G1/13—Traffic control systems for road vehicles indicating the position of vehicles, e.g. scheduled vehicles; Managing passenger vehicles circulating according to a fixed timetable, e.g. buses, trains, trams to a central station ; Indicators in a central station the indicator being in the form of a map
-
- 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
- G10L19/04—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 predictive techniques
- G10L19/26—Pre-filtering or post-filtering
-
- 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
- G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
- G10L25/24—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters the extracted parameters being the cepstrum
-
- 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
- G10L25/27—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique
- G10L25/30—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique using neural networks
-
- 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
- G10L25/45—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of analysis window
-
- 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)
- Multimedia (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Signal Processing (AREA)
- Acoustics & Sound (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Evolutionary Computation (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Alarm Systems (AREA)
Abstract
The invention discloses a kind of monitorings of traffic route sound and exceptional sound recognition system, sound collection end and server end including passing through network connection;Sound collection end includes sound pick-up, sound card, GPS positioning module, data processing module, wireless communication module.Sound pick-up and data processing module can be arranged installation in specified section according to demand;Real-time monitoring is carried out to traffic route sound abnormal sound data are just transmitted to server end and are further identified, volume of transmitted data is greatly reduced when monitoring abnormal sound;The various features value for extracting abnormal sound signal, is identified and is classified in conjunction with neural network;When carrying out identification classification to abnormal sound data, using depth convolutional neural networks (CNN), it is highly suitable for the identification and classification of sound, which can greatly improve training effectiveness and accuracy of identification, the abnormal sound generated in efficient identification traffic route.
Description
Technical field
The present invention relates to sound monitoring and identification technology field, specifically a kind of traffic route sound monitoring and abnormal sound
Identifying system.
Background technique
Abnormal sound refers to that the sound that should not be generated under certain normal environment, the abnormal sound of public place generally include
Explosive sound, impact sound, shriek, gunshot etc..Abnormal sound on traffic route is able to reflect out traffic accident and urgent feelings
The generation of condition passes through the monitoring to traffic route sound, it will be appreciated that the traffic condition of certain road, when being abnormal situation
When, by the identification to abnormal sound, the property of the abnormal conditions can be analyzed.Moreover it is possible to according to monitoring situation judgement
It has the section for being prone to traffic accident or traffic congestion.
Preventing road monitoring system is the important component of intelligent transportation, in existing traffic route monitoring system, usually only
Audio-video monitoring function is without event recognition function, when being abnormal event, usually by artificial playback of monitoring videos into
Row event recognition carries out event recognition by image processing algorithm.Artificial initiative recognition mode is more loaded down with trivial details and time-consuming, figure
As the computationally intensive and computational algorithm that processing identification method needs is complicated.It is therefore, critically important to traffic route progress sound monitoring,
When generating abnormal sound, identification is carried out to judge the type of anomalous event to abnormal sound, is also seemed increasingly important.
Summary of the invention
Present invention aims to overcome that the problem of above-mentioned traffic route monitoring system, propose a kind of traffic route sound
Sound monitoring and exceptional sound recognition system, are acquired sound by sound pick-up and sound card, obtain position by locating module
Abnormal sound data and location information data are sent to server end by wireless network, serviced by information, data processing module
Device end carries out the multi-feature extractions such as MFCC, short-time energy, short-time zero-crossing rate, the characteristic that will be extracted to abnormal sound data
It is input in depth convolutional neural networks and is compared with abnormal sound property data base, the type of abnormal sound finally can be obtained.
By abnormal sound type in conjunction with location information, the position of abnormal sound generation and the class of abnormal sound are showed in map
Type.
To achieve the above object, a kind of traffic route sound monitoring of the present invention and exceptional sound recognition system, including pass through
The sound collection end of network connection and server end;Sound collection end includes sound pick-up, sound card, GPS positioning module, data processing
Module, wireless communication module.
The sound pick-up, for the acquisition to voice signal;
The sound card, for carrying out analog-to-digital conversion to voice signal;
The GPS positioning module, the location information sent for obtaining satellite;
The data processing module judges whether it is abnormal sound for carrying out Preliminary detection to the voice signal of acquisition, and will
The abnormal sound data detected are sent to server end together with location information data;
The wireless communication module, for providing the communication network of transmission data for data processing module;
The server end is mentioned for carrying out the multiple features data such as MFCC, short-time energy, short-time zero-crossing rate to abnormal sound data
It takes, the characteristic of extraction is input in depth convolutional neural networks and is compared with abnormal sound property data base, it is comprehensive
The matching rate of three kinds of features identifies the type of simultaneously output abnormality sound;By abnormal sound type in conjunction with location information, in map
In show abnormal sound generation position and abnormal sound type.
Further, the data processing module judges whether it is abnormal sound to the voice signal progress Preliminary detection of acquisition
Sound, step include:
1) framing handles voice signal as unit of frame;
2) Fast Fourier Transform (FFT) is carried out to voice signal, obtains the corresponding frequency spectrum of each frame;
3) the signal power value of each frame is calculated;
4) decibel value of sound is calculated by the performance number of each frame;
5) decibel value of each frame is made decisions, is then determined as that abnormal sound (according to criteria for noise, is greater than 40db greater than 40db
Then it is determined as noise).
Further, the server end carries out MFCC feature extraction to abnormal sound data, comprising the following steps:
1) preemphasis promotes voice signal high frequency section, signal spectrum is made to become flat;
2) framing handles voice signal as unit of frame;
3) adding window increases a frame data adding window continuity of frame left end and right end;
4) Fast Fourier Transform (FFT) obtains the corresponding frequency spectrum of each frame;
5) Mel is filtered, and the frequency spectrum after Fast Fourier Transform (FFT) is converted to by Mel filter group and embodies human auditory system
Mel frequency spectrum;
6) logarithm is taken, the logarithmic energy of each filter group output is calculated;
7) discrete cosine transform converts logarithmic energy, finds out Mel cepstrum coefficient;
8) dynamic difference parameter is extracted, behavioral characteristics is described with the Difference Spectrum of static nature, effectively improves the identity of system
Energy;
9) the 2nd to the 13rd coefficient and dynamic difference parameter after taking discrete cosine transform are MFCC feature.
Further, the server end carries out short-time energy feature extraction to abnormal sound data, comprising the following steps:
1) framing handles voice signal as unit of frame;
2) adding window increases a frame data adding window continuity of frame left end and right end;
3) it takes absolute value, calculates all sampling point amplitudes in each frame;
4) short-time energy for calculating all sampling points in each frame sums to the short-time energy of all sampling points;
5) the short-time energy value of each frame is taken to correspond to a short-time energy feature.
Further, the server end carries out short-time zero-crossing rate feature extraction, including following step to abnormal sound data
It is rapid:
1) framing handles voice signal as unit of frame;
2) adding window increases a frame data adding window continuity of frame left end and right end;
3) judge whether two sampling points adjacent in each frame have different algebraic symbols, be to occur
Zero passage;
4) number of zero passage in each frame is calculated;
5) number of zero passage in each frame is taken to correspond to a zero-crossing rate feature.
Further, the server end identifies abnormal sound signal using depth convolutional neural networks, including
Following steps:
1) abnormal sound property data base is established;
2) MFCC of the abnormal sound extracted, short-time energy, zero-crossing rate characteristic are input to depth convolutional Neural net
Network;
3) the MFCC characteristic extracted is compared with the MFCC characteristic in database, identifies matching rate by classifier
Highest type;
4) the short-time energy characteristic extracted is compared with the short-time energy characteristic in database, is identified by classifier
The highest type of matching rate out;
5) the short-time zero-crossing rate characteristic extracted is compared with the short-time zero-crossing rate characteristic in database, by classifier
Identify the highest type of matching rate;
6) the in summary comparison matching rate and identification types of three features, obtains best identified result.
Beneficial effects of the present invention: a kind of monitoring of traffic route sound and exceptional sound recognition system, sound pick-up and data
Processing module can be arranged installation in specified section according to demand;Real-time monitoring is carried out to traffic route sound, when monitoring
When abnormal sound, abnormal sound data are just transmitted to server end and are further identified, volume of transmitted data is greatly reduced;
The various features value for extracting abnormal sound signal, is identified and is classified in conjunction with neural network;Abnormal sound data are known
Not Fen Lei when, using depth convolutional neural networks (CNN), be highly suitable for the identification and classification of sound, which can be significantly
Improve training effectiveness and accuracy of identification, the abnormal sound generated in efficient identification traffic route.
Detailed description of the invention
Fig. 1 is traffic route sound monitoring of the present invention and exceptional sound recognition system structure diagram;
Fig. 2 is the schematic diagram at the sound collection end in present system;
Fig. 3 is the schematic diagram of data processing module in present system;
Fig. 4 is the schematic diagram of server-side processes data in present system;
Fig. 5 is the step schematic diagram that MFCC feature is extracted in present system;
Fig. 6 is the step schematic diagram that short-time energy feature is extracted in present system;
Fig. 7 is the step schematic diagram that short-time zero-crossing rate feature is extracted in present system.
Specific embodiment
The content of present invention is further described below with reference to embodiment and attached drawing, but is not limitation of the invention.
Embodiment
As shown in Fig. 1 system structure diagram, traffic route sound of the invention, which is monitored with exceptional sound recognition system, includes
Sound collection end 7 and server end 6 by network connection, sound collection end 7 include sound pick-up 1, sound card 2, GPS positioning module
3, data processing module 4, wireless communication module 5.
Sound pick-up 1 acquires and amplifies voice signal, by transmission of sound signals to sound card 2.Sound card 2 believes collected sound
Number (analog signal) is converted into digital signal, by digital data transmission to data processing module 4.GPS positioning module 3 receives satellite
Location information is transmitted to data processing module 4 by the location information sent.Data processing module 4 is to the voice signal received
Preliminary detection is carried out, judges whether there is abnormal sound, module 5 is sent to clothes by wireless communication if detecting abnormal sound
Business device end 6, while the GPS positioning information received is handled, select two-dimensional position information data mould by wireless communication
Block 5 is sent to server end 6.Server end 6 carries out multi-feature extraction to the abnormal sound data received, in conjunction with depth convolution mind
Classification and Identification is carried out through network.It is final to combine the location information received, position and classification recognition result are occurred into for abnormal sound on ground
It is presented on figure.
As shown in Fig. 2, microphone collected sound signal, this voice signal is analog signal in sound pick-up 1, pass through operation
Amplifier carries out first order amplification, then carries out second level amplification, mould of the output by amplification by automatic gain amplifier (AGC)
Quasi- signal.The twin-stage amplifying circuit efficiently controls the power of voice signal, avoids moment high-decibel sound to subsequent equipment
It influences.In sound card 2, by the analog signal of amplification after over-sampling, quantization, coding, output data processing module 4 is identifiable
PCM digital signal.
If Fig. 3 data processing module 4 is handled shown in schematic diagram data, data processing module 4 is first to the sound received
Signal carries out framing, and signal is handled as unit of frame, then carries out Fast Fourier Transform (FFT) to each frame signal
(FFT), the corresponding frequency spectrum of each frame is obtained, the signal power value of each frame is then calculated, is calculated by the performance number of each frame
The decibel value of sound out finally makes decisions the decibel value of each frame, is then determined as abnormal sound (according to noise greater than 40db
Standard is then determined as noise greater than 40db).It particularly illustrates, if an exception occurs sound, then the voice signal is several continuous
The decibel value of frame is all greater than 40db.Meanwhile the extraction of two-dimensional position coordinate is carried out to GPS positioning information.The abnormal sound that will be extracted
Sound data and two-dimensional position coordinate data are packaged, the network transmission that module 5 provides by wireless communication to server terminal.
As shown in Fig. 4 server-side processes schematic diagram data, server end 6 receives the number that data processing module 4 is sent
According to packet, voice data and two-dimensional position coordinate are obtained by unpacking.For voice data, saved first to local, then to sound
Sound data are filtered, and then carry out multi-feature extraction to voice signal, and characteristic is finally inputted depth convolutional Neural net
Network is identified and is classified, and obtains classification results.For classification results, it is saved to corresponding library and carries out abnormal sound database
It establishes, such as saves traffic accident impact sound to traffic accident impact sound database.For two-dimensional position coordinate, first preservation to local, so
Carry out coordinate conversion afterwards, earth latitude and longitude coordinates be converted into corresponding map reference, the embodiment of the present invention call Baidu/
Google Maps.Finally, position and classification that abnormal sound occurs is presented in combining classification result and position coordinates on map.
Further, as described above, being related to carrying out multi-feature extraction to voice signal.Abnormal sound be it is a kind of it is aperiodic,
The random signal of non-stationary only cannot fully be described abnormal sound with a kind of feature in time domain or frequency domain.?
In short time, generally 10ms-30ms, abnormal sound signal can be considered a kind of short-term stationarity signal, be based on this characteristic, can extract
Multiple features in voice signal time-domain and frequency-domain, are identified using multiple features, discrimination can be improved.To abnormal sound message
Number carry out multi-feature extraction, including three MFCC, short-time energy, short-time zero-crossing rate features.
Further, as shown in figure 5, MFCC feature extraction the following steps are included:
1) preemphasis promotes voice signal high frequency section, signal spectrum is made to become flat;
2) framing handles voice signal as unit of frame;
3) adding window increases a frame data adding window continuity of frame left end and right end;
4) Fast Fourier Transform (FFT) obtains the corresponding frequency spectrum of each frame;
5) Mel is filtered, and the frequency spectrum after Fast Fourier Transform (FFT) is converted to by Mel filter group and embodies human auditory system
Mel frequency spectrum;
6) logarithm is taken, the logarithmic energy of each filter group output is calculated;
7) discrete cosine transform converts logarithmic energy, finds out Mel cepstrum coefficient;
8) dynamic difference parameter is extracted, behavioral characteristics is described with the Difference Spectrum of static nature, effectively improves the identity of system
Energy;
9) the 2nd to the 13rd coefficient and dynamic difference parameter after taking discrete cosine transform are MFCC feature.
Further, as shown in fig. 6, short-time energy feature extraction the following steps are included:
1) framing handles voice signal as unit of frame;
2) adding window increases a frame data adding window continuity of frame left end and right end;
3) it takes absolute value, calculates all sampling point amplitudes in each frame;
4) short-time energy for calculating all sampling points in each frame sums to the short-time energy of all sampling points;
5) the short-time energy value of each frame is taken to correspond to a short-time energy feature.
Further, as shown in fig. 7, short-time zero-crossing rate feature extraction the following steps are included:
1) framing handles voice signal as unit of frame;
2) adding window increases a frame data adding window continuity of frame left end and right end;
3) judge whether two sampling points adjacent in each frame have different algebraic symbols, be that zero passage has occurred;
4) number of zero passage in each frame is calculated;
5) number of zero passage in each frame is taken to correspond to a zero-crossing rate feature.
Further, it is related to identification of the depth convolutional neural networks to abnormal sound signal, comprising the following steps:
1) abnormal sound property data base is established;
2) MFCC of the abnormal sound extracted, short-time energy, zero-crossing rate characteristic are input to depth convolutional neural networks;
3) the MFCC characteristic extracted is compared with the MFCC characteristic in database, identifies matching rate by classifier
Highest type;
4) the short-time energy characteristic extracted is compared with the short-time energy characteristic in database, is identified by classifier
The highest type of matching rate out;
5) the short-time zero-crossing rate characteristic extracted is compared with the short-time zero-crossing rate characteristic in database, by classifier
Identify the highest type of matching rate;
6) the in summary comparison matching rate and identification types of three features, obtains best identified result.
Claims (6)
1. a kind of traffic route sound monitoring and exceptional sound recognition system, it is characterised in that: including the sound by network connection
Sound collection terminal and server end;
Sound collection end includes sound pick-up, sound card, GPS positioning module, data processing module, wireless communication module;
The sound pick-up, for the acquisition to voice signal;
The sound card, for carrying out analog-to-digital conversion to voice signal;
The GPS positioning module, the location information sent for obtaining satellite;
The data processing module judges whether it is abnormal sound for carrying out Preliminary detection to the voice signal of acquisition, and will
The abnormal sound data detected are sent to server end together with location information data;
The wireless communication module, for providing the communication network of transmission data for data processing module;
The server end, for being mentioned to three kinds of abnormal sound data progress MFCC, short-time energy, short-time zero-crossing rate characteristics
It takes, the characteristic of extraction is input in depth convolutional neural networks and is compared with abnormal sound property data base, it is comprehensive
The matching rate of three kinds of features identifies the type of simultaneously output abnormality sound;By abnormal sound type in conjunction with location information, in map
In show abnormal sound generation position and abnormal sound type.
2. the monitoring of traffic route sound and exceptional sound recognition system according to claim 1, it is characterised in that: the number
Preliminary detection is carried out according to voice signal of the processing module to acquisition and judges whether it is abnormal sound, and step includes:
1) framing handles voice signal as unit of frame;
2) Fast Fourier Transform (FFT) is carried out to voice data, obtains the corresponding frequency spectrum of each frame;
3) the signal power value of each frame is calculated;
4) decibel value of sound is calculated by the performance number of each frame;
5) decibel value of each frame is made decisions, is then determined as abnormal sound greater than 40db.
3. the monitoring of traffic route sound and exceptional sound recognition system according to claim 1, it is characterised in that: the service
Device end carries out MFCC feature extraction to abnormal sound data, comprising the following steps:
1) preemphasis, promotion signal high frequency section make signal spectrum become flat;
2) framing handles signal as unit of frame;
3) adding window increases a frame data adding window continuity of frame left end and right end;
4) Fast Fourier Transform (FFT) obtains the corresponding frequency spectrum of each frame;
5) Mel is filtered, and the frequency spectrum after Fast Fourier Transform (FFT) is converted to by Mel filter group and embodies human auditory system
Mel frequency spectrum;
6) logarithm is taken, the logarithmic energy of each filter group output is calculated;
7) discrete cosine transform converts logarithmic energy, finds out Mel cepstrum coefficient;
8) dynamic difference parameter is extracted, behavioral characteristics is described with the Difference Spectrum of static nature, effectively improves the identity of system
Energy;
9) the 2nd to the 13rd coefficient and dynamic difference parameter after taking discrete cosine transform are MFCC feature.
4. the monitoring of traffic route sound and exceptional sound recognition system according to claim 1, it is characterised in that: the service
Device end carries out short-time energy feature extraction to abnormal sound data, comprising the following steps:
1) framing handles signal as unit of frame;
2) adding window increases a frame data adding window continuity of frame left end and right end;
3) it takes absolute value, calculates all sampling point amplitudes in each frame;
4) short-time energy for calculating all sampling points in each frame sums to the short-time energy of all sampling points;
5) the short-time energy value of each frame is taken to correspond to a short-time energy feature.
5. the monitoring of traffic route sound and exceptional sound recognition system according to claim 1, it is characterised in that: the service
Device end carries out zero-crossing rate feature extraction to abnormal sound data, comprising the following steps:
1) framing handles signal as unit of frame;
2) adding window increases a frame data adding window continuity of frame left end and right end;
3) judge whether two sampling points adjacent in each frame have different algebraic symbols, be that zero passage has occurred;
4) number of zero passage in each frame is calculated;
5) number of zero passage in each frame is taken to correspond to a zero-crossing rate feature.
6. the monitoring of traffic route sound and exceptional sound recognition system according to claim 1, it is characterised in that: the service
Device end identifies abnormal sound signal using depth convolutional neural networks, comprising the following steps:
1) abnormal sound property data base is established;
2) MFCC of the abnormal sound extracted, short-time energy, zero-crossing rate characteristic are input to depth convolutional neural networks;
3) the MFCC characteristic extracted is compared with the MFCC characteristic in database, identifies matching rate by classifier
Highest type;
4) the short-time energy characteristic extracted is compared with the short-time energy characteristic in database, is identified by classifier
The highest type of matching rate out;
5) the short-time zero-crossing rate characteristic extracted is compared with the short-time zero-crossing rate characteristic in database, by classifier
Identify the highest type of matching rate;
6) the in summary comparison matching rate and identification types of three features, obtains best identified result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810851609.8A CN109087655A (en) | 2018-07-30 | 2018-07-30 | A kind of monitoring of traffic route sound and exceptional sound recognition system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810851609.8A CN109087655A (en) | 2018-07-30 | 2018-07-30 | A kind of monitoring of traffic route sound and exceptional sound recognition system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109087655A true CN109087655A (en) | 2018-12-25 |
Family
ID=64833348
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810851609.8A Pending CN109087655A (en) | 2018-07-30 | 2018-07-30 | A kind of monitoring of traffic route sound and exceptional sound recognition system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109087655A (en) |
Cited By (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109615867A (en) * | 2019-01-28 | 2019-04-12 | 大连海事大学 | A kind of Intelligent road system towards traffic dispersion |
CN109658943A (en) * | 2019-01-23 | 2019-04-19 | 平安科技(深圳)有限公司 | A kind of detection method of audio-frequency noise, device, storage medium and mobile terminal |
CN109767785A (en) * | 2019-03-06 | 2019-05-17 | 河北工业大学 | Ambient noise method for identifying and classifying based on convolutional neural networks |
CN109785857A (en) * | 2019-02-28 | 2019-05-21 | 桂林电子科技大学 | Abnormal sound event recognition method based on MFCC+MP fusion feature |
CN109784411A (en) * | 2019-01-23 | 2019-05-21 | 四川虹微技术有限公司 | To the defence method of resisting sample, device, system and storage medium |
CN109948739A (en) * | 2019-04-22 | 2019-06-28 | 桂林电子科技大学 | Ambient sound event acquisition and Transmission system based on support vector machines |
CN110164472A (en) * | 2019-04-19 | 2019-08-23 | 天津大学 | Noise classification method based on convolutional neural networks |
CN110176248A (en) * | 2019-05-23 | 2019-08-27 | 广西交通科学研究院有限公司 | Road sound identification method, system, computer equipment and readable storage medium storing program for executing |
CN110706721A (en) * | 2019-10-17 | 2020-01-17 | 南京林业大学 | Electric precipitation spark discharge identification method based on BP neural network |
CN110718235A (en) * | 2019-09-20 | 2020-01-21 | 精锐视觉智能科技(深圳)有限公司 | Abnormal sound detection method, electronic device and storage medium |
CN110890102A (en) * | 2019-09-07 | 2020-03-17 | 创新奇智(重庆)科技有限公司 | Engine defect detection algorithm based on RNN voiceprint recognition |
CN111009261A (en) * | 2019-12-10 | 2020-04-14 | Oppo广东移动通信有限公司 | Arrival reminding method, device, terminal and storage medium |
CN111127876A (en) * | 2019-11-18 | 2020-05-08 | 腾讯科技(深圳)有限公司 | Information extraction method and device for Internet of vehicles |
CN111341334A (en) * | 2020-03-06 | 2020-06-26 | 东莞理工学院 | Noise reduction and abnormal sound detection system and method applied to rail transit |
CN111370027A (en) * | 2020-03-02 | 2020-07-03 | 乐鑫信息科技(上海)股份有限公司 | Off-line embedded abnormal sound detection system and method |
CN112866639A (en) * | 2021-01-07 | 2021-05-28 | 北京家人智能科技有限公司 | Patrol warning method and device |
CN113074967A (en) * | 2020-01-06 | 2021-07-06 | 北京谛声科技有限责任公司 | Abnormal sound detection method and device, storage medium and electronic equipment |
CN113077634A (en) * | 2021-03-19 | 2021-07-06 | 上海电机学院 | Method for assisting traffic monitoring |
CN113345235A (en) * | 2021-06-07 | 2021-09-03 | 恒明星光智慧文化科技(深圳)有限公司 | Road intelligence emergency treatment device, sculpture, wisdom street lamp |
CN113539298A (en) * | 2021-07-19 | 2021-10-22 | 中通服咨询设计研究院有限公司 | Sound big data analysis calculates imaging system based on cloud limit end |
CN115116230A (en) * | 2022-07-26 | 2022-09-27 | 浪潮卓数大数据产业发展有限公司 | Traffic environment monitoring method, equipment and medium |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6105015A (en) * | 1997-02-03 | 2000-08-15 | The United States Of America As Represented By The Secretary Of The Navy | Wavelet-based hybrid neurosystem for classifying a signal or an image represented by the signal in a data system |
CN102148032A (en) * | 2010-12-03 | 2011-08-10 | 北京声迅电子有限公司 | Abnormal sound detection method and system for ATM (Automatic Teller Machine) |
CN102737480A (en) * | 2012-07-09 | 2012-10-17 | 广州市浩云安防科技股份有限公司 | Abnormal voice monitoring system and method based on intelligent video |
CN103198838A (en) * | 2013-03-29 | 2013-07-10 | 苏州皓泰视频技术有限公司 | Abnormal sound monitoring method and abnormal sound monitoring device used for embedded system |
CN105810213A (en) * | 2014-12-30 | 2016-07-27 | 浙江大华技术股份有限公司 | Typical abnormal sound detection method and device |
CN107086036A (en) * | 2017-04-19 | 2017-08-22 | 杭州派尼澳电子科技有限公司 | A kind of freeway tunnel method for safety monitoring |
US20180080812A1 (en) * | 2017-07-25 | 2018-03-22 | University Of Electronic Science And Technology Of China | Distributed optical fiber sensing signal processing method for safety monitoring of underground pipe network |
-
2018
- 2018-07-30 CN CN201810851609.8A patent/CN109087655A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6105015A (en) * | 1997-02-03 | 2000-08-15 | The United States Of America As Represented By The Secretary Of The Navy | Wavelet-based hybrid neurosystem for classifying a signal or an image represented by the signal in a data system |
CN102148032A (en) * | 2010-12-03 | 2011-08-10 | 北京声迅电子有限公司 | Abnormal sound detection method and system for ATM (Automatic Teller Machine) |
CN102737480A (en) * | 2012-07-09 | 2012-10-17 | 广州市浩云安防科技股份有限公司 | Abnormal voice monitoring system and method based on intelligent video |
CN103198838A (en) * | 2013-03-29 | 2013-07-10 | 苏州皓泰视频技术有限公司 | Abnormal sound monitoring method and abnormal sound monitoring device used for embedded system |
CN105810213A (en) * | 2014-12-30 | 2016-07-27 | 浙江大华技术股份有限公司 | Typical abnormal sound detection method and device |
CN107086036A (en) * | 2017-04-19 | 2017-08-22 | 杭州派尼澳电子科技有限公司 | A kind of freeway tunnel method for safety monitoring |
US20180080812A1 (en) * | 2017-07-25 | 2018-03-22 | University Of Electronic Science And Technology Of China | Distributed optical fiber sensing signal processing method for safety monitoring of underground pipe network |
Non-Patent Citations (1)
Title |
---|
苏健民 等: "基于语音信号端点检测技术的研究应用", 《林业机械与木工设备》 * |
Cited By (30)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109658943A (en) * | 2019-01-23 | 2019-04-19 | 平安科技(深圳)有限公司 | A kind of detection method of audio-frequency noise, device, storage medium and mobile terminal |
CN109784411A (en) * | 2019-01-23 | 2019-05-21 | 四川虹微技术有限公司 | To the defence method of resisting sample, device, system and storage medium |
CN109615867A (en) * | 2019-01-28 | 2019-04-12 | 大连海事大学 | A kind of Intelligent road system towards traffic dispersion |
CN109785857A (en) * | 2019-02-28 | 2019-05-21 | 桂林电子科技大学 | Abnormal sound event recognition method based on MFCC+MP fusion feature |
CN109785857B (en) * | 2019-02-28 | 2020-08-14 | 桂林电子科技大学 | Abnormal sound event identification method based on MFCC + MP fusion characteristics |
CN109767785A (en) * | 2019-03-06 | 2019-05-17 | 河北工业大学 | Ambient noise method for identifying and classifying based on convolutional neural networks |
CN110164472A (en) * | 2019-04-19 | 2019-08-23 | 天津大学 | Noise classification method based on convolutional neural networks |
CN109948739A (en) * | 2019-04-22 | 2019-06-28 | 桂林电子科技大学 | Ambient sound event acquisition and Transmission system based on support vector machines |
CN110176248A (en) * | 2019-05-23 | 2019-08-27 | 广西交通科学研究院有限公司 | Road sound identification method, system, computer equipment and readable storage medium storing program for executing |
CN110176248B (en) * | 2019-05-23 | 2020-12-22 | 广西交科集团有限公司 | Road voice recognition method, system, computer device and readable storage medium |
CN110890102A (en) * | 2019-09-07 | 2020-03-17 | 创新奇智(重庆)科技有限公司 | Engine defect detection algorithm based on RNN voiceprint recognition |
CN110718235A (en) * | 2019-09-20 | 2020-01-21 | 精锐视觉智能科技(深圳)有限公司 | Abnormal sound detection method, electronic device and storage medium |
CN110718235B (en) * | 2019-09-20 | 2022-07-01 | 精锐视觉智能科技(深圳)有限公司 | Abnormal sound detection method, electronic device and storage medium |
CN110706721A (en) * | 2019-10-17 | 2020-01-17 | 南京林业大学 | Electric precipitation spark discharge identification method based on BP neural network |
CN111127876A (en) * | 2019-11-18 | 2020-05-08 | 腾讯科技(深圳)有限公司 | Information extraction method and device for Internet of vehicles |
CN111127876B (en) * | 2019-11-18 | 2021-11-05 | 腾讯科技(深圳)有限公司 | Information extraction method and device for Internet of vehicles |
CN111009261A (en) * | 2019-12-10 | 2020-04-14 | Oppo广东移动通信有限公司 | Arrival reminding method, device, terminal and storage medium |
CN111009261B (en) * | 2019-12-10 | 2022-11-15 | Oppo广东移动通信有限公司 | Arrival reminding method, device, terminal and storage medium |
CN113074967B (en) * | 2020-01-06 | 2022-12-16 | 北京谛声科技有限责任公司 | Abnormal sound detection method and device, storage medium and electronic equipment |
CN113074967A (en) * | 2020-01-06 | 2021-07-06 | 北京谛声科技有限责任公司 | Abnormal sound detection method and device, storage medium and electronic equipment |
CN111370027A (en) * | 2020-03-02 | 2020-07-03 | 乐鑫信息科技(上海)股份有限公司 | Off-line embedded abnormal sound detection system and method |
CN111370027B (en) * | 2020-03-02 | 2023-04-07 | 乐鑫信息科技(上海)股份有限公司 | Off-line embedded abnormal sound detection system and method |
CN111341334A (en) * | 2020-03-06 | 2020-06-26 | 东莞理工学院 | Noise reduction and abnormal sound detection system and method applied to rail transit |
CN112866639A (en) * | 2021-01-07 | 2021-05-28 | 北京家人智能科技有限公司 | Patrol warning method and device |
CN112866639B (en) * | 2021-01-07 | 2023-04-28 | 珠海市横琴盈实科技研发有限公司 | Patrol warning method and equipment |
CN113077634A (en) * | 2021-03-19 | 2021-07-06 | 上海电机学院 | Method for assisting traffic monitoring |
CN113345235A (en) * | 2021-06-07 | 2021-09-03 | 恒明星光智慧文化科技(深圳)有限公司 | Road intelligence emergency treatment device, sculpture, wisdom street lamp |
CN113539298A (en) * | 2021-07-19 | 2021-10-22 | 中通服咨询设计研究院有限公司 | Sound big data analysis calculates imaging system based on cloud limit end |
CN113539298B (en) * | 2021-07-19 | 2023-11-14 | 中通服咨询设计研究院有限公司 | Sound big data analysis and calculation imaging system based on cloud edge end |
CN115116230A (en) * | 2022-07-26 | 2022-09-27 | 浪潮卓数大数据产业发展有限公司 | Traffic environment monitoring method, equipment and medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109087655A (en) | A kind of monitoring of traffic route sound and exceptional sound recognition system | |
KR100636317B1 (en) | Distributed Speech Recognition System and method | |
KR101496067B1 (en) | Method and apparatus for determining location of mobile device | |
JP6466334B2 (en) | Real-time traffic detection | |
KR101250668B1 (en) | Method for recogning emergency speech using gmm | |
CN110600008A (en) | Voice wake-up optimization method and system | |
CN111312286A (en) | Age identification method, age identification device, age identification equipment and computer readable storage medium | |
WO2017045429A1 (en) | Audio data detection method and system and storage medium | |
CN113947376B (en) | C/S (computer/subscriber line) card punching method and device based on multiple biological characteristics | |
CN103514877A (en) | Vibration signal characteristic parameter extracting method | |
CN110800053A (en) | Method and apparatus for obtaining event indications based on audio data | |
CN105139481A (en) | Vehicle speaker recognition system | |
KR20190046569A (en) | Acoustic Tunnel Accident Detection System | |
CN115510265A (en) | Method and system for judging animal hazard distribution of pole tower in power transmission line | |
CN113793624A (en) | Acoustic scene classification method | |
CN110580915B (en) | Sound source target identification system based on wearable equipment | |
CN112233696A (en) | Oil field pumping unit abnormal sound detection and reporting system based on artificial intelligence and big data | |
CN114743562B (en) | Method and system for recognizing airplane voiceprint, electronic equipment and storage medium | |
CN113472466B (en) | Black broadcast monitoring system based on emergency broadcast system | |
CN107548007B (en) | Detection method and device of audio signal acquisition equipment | |
EP3309777A1 (en) | Device and method for audio frame processing | |
CN114694682A (en) | Method, system and device for detecting abnormity of brake system | |
Nicolae et al. | A Method for chainsaw sound detection based on Haar-like features | |
Grama et al. | About quantization of audio signals for wildlife intruder detection systems | |
CN117636909B (en) | Data processing method, device, equipment and computer readable storage medium |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181225 |
|
RJ01 | Rejection of invention patent application after publication |