CN111540347A - Cable tunnel monitoring method and system based on audio - Google Patents
Cable tunnel monitoring method and system based on audio Download PDFInfo
- Publication number
- CN111540347A CN111540347A CN202010399900.3A CN202010399900A CN111540347A CN 111540347 A CN111540347 A CN 111540347A CN 202010399900 A CN202010399900 A CN 202010399900A CN 111540347 A CN111540347 A CN 111540347A
- Authority
- CN
- China
- Prior art keywords
- audio
- sound
- cable tunnel
- sound source
- signal
- 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
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000012544 monitoring process Methods 0.000 title claims abstract description 26
- 230000005236 sound signal Effects 0.000 claims abstract description 45
- 238000012545 processing Methods 0.000 claims abstract description 26
- 238000001228 spectrum Methods 0.000 claims description 18
- 230000003321 amplification Effects 0.000 claims description 12
- 238000001914 filtration Methods 0.000 claims description 12
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 12
- 238000012360 testing method Methods 0.000 claims description 11
- 238000012549 training Methods 0.000 claims description 10
- 238000005070 sampling Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 5
- 239000011159 matrix material Substances 0.000 claims description 5
- 229910052760 oxygen Inorganic materials 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000001514 detection method Methods 0.000 claims description 4
- 238000013139 quantization Methods 0.000 claims description 4
- 229910052717 sulfur Inorganic materials 0.000 claims description 4
- 238000012546 transfer Methods 0.000 claims description 4
- 230000008569 process Effects 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 230000004807 localization Effects 0.000 description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 3
- 238000013461 design Methods 0.000 description 2
- 238000009432 framing Methods 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000008054 signal transmission Effects 0.000 description 2
- UGFAIRIUMAVXCW-UHFFFAOYSA-N Carbon monoxide Chemical compound [O+]#[C-] UGFAIRIUMAVXCW-UHFFFAOYSA-N 0.000 description 1
- RWSOTUBLDIXVET-UHFFFAOYSA-N Dihydrogen sulfide Chemical compound S RWSOTUBLDIXVET-UHFFFAOYSA-N 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000004204 blood vessel Anatomy 0.000 description 1
- 229910002091 carbon monoxide Inorganic materials 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 239000007789 gas Substances 0.000 description 1
- 229910000037 hydrogen sulfide Inorganic materials 0.000 description 1
- 230000009545 invasion Effects 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- 239000001301 oxygen Substances 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000007480 spreading Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L15/14—Speech classification or search using statistical models, e.g. Hidden Markov Models [HMMs]
- G10L15/142—Hidden Markov Models [HMMs]
- G10L15/144—Training of HMMs
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01H—MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC OR INFRASONIC WAVES
- G01H17/00—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves, not provided for in the preceding groups
-
- 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
- G10L25/18—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 spectral information of each sub-band
-
- 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
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Computational Linguistics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- General Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Probability & Statistics with Applications (AREA)
- Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)
Abstract
The invention discloses a cable tunnel monitoring method and a system based on audio, wherein the method comprises the following steps: s1, collecting audio in the cable tunnel; s2, processing the collected audio signal; s3, extracting characteristic values of the collected audio signals; s4, identifying faults according to the extracted features; and S5, positioning the fault. The invention can analyze the sound in the cable tunnel, can analyze and position the sound source, and helps workers to intuitively analyze the operation condition in the tunnel.
Description
Technical Field
The invention relates to a cable tunnel monitoring method and system based on audio frequency, and belongs to the technical field of cable tunnels.
Background
With the annual increase of the investment of national power infrastructure, the length of a power cable tunnel is also rapidly increasing, and the operation of the power cable tunnel faces huge pressure because the increase speed of operation maintenance personnel is far from the increase speed of the power infrastructure.
The application of the cable tunnel meets the power supply requirements of urban high load density and cabling, can effectively guarantee the transmission capacity of the power channel, and simultaneously improves the utilization rate of channel resources. The safety of the power cable tunnel is concerned with various aspects of cities, the power cable tunnel not only needs to monitor a cable body, such as a cable joint, but also needs to monitor the environment in the cable tunnel, such as oxygen, hydrogen sulfide, carbon monoxide, methane and other harmful gases, and meanwhile, the cable tunnel is a blood vessel of the city and is used for conveying continuous energy for the city.
How to guarantee that cable in the tunnel does not influence power cable operation incident because of ponding, catching fire that the condition such as overload, overheat lead to and in time monitor the phenomenon that someone in the tunnel or well lid open etc. probably lead to power cable and send the accident, it is the urgent in priority to adopt modernized technical means to improve cable tunnel operation maintenance level.
Disclosure of Invention
In order to solve the problems, the invention provides a cable tunnel monitoring method and system based on audio frequency, which can monitor the fault of sound generated in the cable tunnel and avoid the fault spreading and expansion.
The technical scheme adopted for solving the technical problems is as follows:
on one hand, the embodiment of the invention provides an audio-based cable tunnel monitoring method, which comprises the following steps:
s1, collecting audio in the cable tunnel;
s2, processing the collected audio signal;
s3, extracting characteristic values of the collected audio signals;
s4, identifying faults according to the extracted features;
and S5, positioning the fault.
As a possible implementation manner of this embodiment, in step S1, the audio frequency in the cable tunnel is acquired by using a plurality of capacitive test microphones, and the plurality of capacitive test microphones are disposed at respective detection points in the cable tunnel and constitute a microphone matrix.
As a possible implementation manner of this embodiment, the step S2 specifically includes:
carrying out two-stage in-phase proportional amplification processing on the collected sound signals;
filtering the amplified sound signal;
and sampling the sound signal after the filtering processing.
As a possible implementation manner of this embodiment, the step S3 specifically includes:
the audio signal S (n) is preprocessed to obtain a time domain signal X (n) of each audio frame, and the time domain signal is processed by FFT to obtain a linear spectrum X (k).
Performing square operation on the amplitude of the linear spectrum X (k), and obtaining an energy spectrum; the energy spectrum is band-pass filtered.
The transfer function of the band pass filter is:
the logarithmic energy calculation formula output by each filter bank is as follows:
obtaining a voice characteristic parameter MFCC through DCT (discrete cosine transform):
as a possible implementation manner of this embodiment, the step S4 specifically includes:
the extracted features are a set of random vector sequences O ═ O1o2,…oT；
The vector O is O by vector quantization method1o2,…oTAdjusting the sequence to be an observation sequence needing to be input in a hidden Markov model, and obtaining hidden Markov models of different sound signals for the sequence through a hidden Markov model training algorithm to form a fault model library;
the fault sound signal to be identified is input, and O is obtained by the previous method1o2,…oTThen, the output probability P (O | M) of each hidden Markov model is calculated, and the digital type corresponding to the hidden Markov model with the maximum P (O | M) is the recognition result.
As a possible implementation manner of this embodiment, the hidden markov model training algorithm specifically includes:
a hidden markov model is described by M ═ { S, O, a, B, pi, F }, where S is the set of all states in the HMM, O is the set of observed value symbols that are output, and a ═ { a ═ a }ijI, j is 1,2, …, N }, B is a set of output observation value probabilities, pi is a set of system initial state probabilities, and F is a set of system termination states;
segmenting the audio of the same type, wherein the length of the intercepted audio of different types is 20 frames to 500 frames;
setting initial values of pi and A, B at will, obtaining the hidden state of the audio frequency according to the Viterbi algorithm, and calculating the probability of each observation value of each state according to the estimated hidden state and the observation value to obtain the initial value of B.
As a possible implementation manner of this embodiment, the step S4 specifically includes: and calculating the specific position of the sound source according to the time difference from the audio sound source to the audio acquisition equipment.
As a possible implementation manner of this embodiment, the calculating a specific position of the sound source according to a time difference from the audio sound source to the audio acquisition device specifically includes:
setting the coordinates of the audio acquisition equipment of the quaternary array as M1(d/2,0,0), M2(0, d/2,0), M3(-d/2,0,0), M4(0, -d/2,0), M1, M2, M3 and M4 as array elements of the audio acquisition equipment respectively, the coordinates of a sound source as (x, y, z), the distance from the sound source to the origin of coordinates as r, the elevation angle as theta, the azimuth angle as phi and d as the interval of the array elements;
relative to M1, the time delay of sound source arriving at array elements M2, M3 and M4 is tau21，τ31，τ41The sound differences from the sound source to M2, M3 and M4 are d21, d31 and d41 respectively, and the position of the sound source is solved according to the following equation:
solving to obtain the sound source position as follows:
as a possible implementation manner of this embodiment, the audio in the cable tunnel includes a water dropping sound audio, a fire catching sound audio, a cable joint failure sound audio, a tunnel person walking sound audio, a manhole cover opening sound audio, and a vehicle pressure manhole cover sound audio.
On the other hand, the cable tunnel monitoring system based on audio provided by the embodiment of the present invention includes:
the signal acquisition module is used for acquiring sound in the cable tunnel;
the signal processing module is used for processing the acquired audio signals;
the characteristic extraction module is used for extracting a characteristic value from the acquired audio signal;
the fault identification module is used for identifying faults according to the extracted features;
and the fault positioning module is used for positioning the fault.
The technical scheme of the embodiment of the invention has the following beneficial effects:
according to the technical scheme of the embodiment of the invention, the sound of the cable tunnel can be analyzed, the sound source signal can be positioned, on one hand, the working personnel can be helped to predict the fault in the cable tunnel, and on the other hand, the video acquisition can be carried out on the sound source signal by combining a video monitoring system in the tunnel, so that the anti-theft invasion in the tunnel can be identified.
The technical scheme of the embodiment of the invention can analyze the sound in the cable tunnel, can analyze and position the sound source, and helps workers to intuitively analyze the operation condition in the tunnel.
The cable tunnel monitoring system based on the audio frequency adopts a modular design, and is simple in structure and easy to realize.
Description of the drawings:
FIG. 1 is a flow diagram illustrating a method of audio-based cable tunnel monitoring in accordance with an exemplary embodiment;
FIG. 2 is a block diagram illustrating an audio-based cable tunnel monitoring system in accordance with an exemplary embodiment;
FIG. 3 is a flow diagram illustrating a sound source localization process according to an exemplary embodiment.
Detailed Description
The invention is further illustrated by the following examples in conjunction with the accompanying drawings:
in order to clearly explain the technical features of the present invention, the following detailed description of the present invention is provided with reference to the accompanying drawings. The following disclosure provides many different embodiments, or examples, for implementing different features of the invention. To simplify the disclosure of the present invention, the components and arrangements of specific examples are described below. Furthermore, the present invention may repeat reference numerals and/or letters in the various examples. This repetition is for the purpose of simplicity and clarity and does not in itself dictate a relationship between the various embodiments and/or configurations discussed. It should be noted that the components illustrated in the figures are not necessarily drawn to scale. Descriptions of well-known components and processing techniques and procedures are omitted so as to not unnecessarily limit the invention.
Fig. 1 is a flow chart illustrating a method of audio-based cable tunnel monitoring according to an exemplary embodiment. As shown in fig. 1, an audio-based cable tunnel monitoring method provided in an embodiment of the present invention includes the following steps:
s1, collecting audio in the cable tunnel;
s2, processing the collected audio signal;
s3, extracting characteristic values of the collected audio signals;
s4, identifying faults according to the extracted features;
and S5, positioning the fault.
As a possible implementation manner of this embodiment, in step S1, the audio frequency in the cable tunnel is acquired by using a plurality of capacitive test microphones, and the plurality of capacitive test microphones are disposed at respective detection points in the cable tunnel and constitute a microphone matrix.
As a possible implementation manner of this embodiment, the step S2 specifically includes:
carrying out two-stage in-phase proportional amplification processing on the collected sound signals;
filtering the amplified sound signal;
and sampling the sound signal after the filtering processing.
As a possible implementation manner of this embodiment, the step S3 specifically includes:
the audio signal S (n) is preprocessed to obtain a time domain signal X (n) of each audio frame, and the time domain signal is processed by FFT to obtain a linear spectrum X (k).
Performing square operation on the amplitude of the linear spectrum X (k), and obtaining an energy spectrum; the energy spectrum is band-pass filtered.
The transfer function of the band pass filter is:
the logarithmic energy calculation formula output by each filter bank is as follows:
obtaining a voice characteristic parameter MFCC through DCT (discrete cosine transform):
as a possible implementation manner of this embodiment, the step S4 specifically includes:
the extracted features are a set of random vector sequences O ═ O1o2,…oT；
The vector O is O by vector quantization method1o2,…oTAdjusting the sequence to be an observation sequence needing to be input in a hidden Markov model, and obtaining hidden Markov models of different sound signals for the sequence through a hidden Markov model training algorithm to form a fault model library;
the fault sound signal to be identified is input, and O is obtained by the previous method1o2,…oTThen, the output probability P (O | M) of each hidden Markov model is calculated, and the digital type corresponding to the hidden Markov model with the maximum P (O | M) is the recognition result.
As a possible implementation manner of this embodiment, the hidden markov model training algorithm specifically includes:
a hidden markov model is described by M ═ { S, O, a, B, pi, F }, where S is the set of all states in the HMM, O is the set of observed value symbols that are output, and a ═ { a ═ a }ijI, j is 1,2, …, N }, B is a set of output observation value probabilities, pi is a set of system initial state probabilities, and F is a set of system termination states;
segmenting the audio of the same type, wherein the length of the intercepted audio of different types is 20 frames to 500 frames;
setting initial values of pi and A, B at will, obtaining the hidden state of the audio frequency according to the Viterbi algorithm, and calculating the probability of each observation value of each state according to the estimated hidden state and the observation value to obtain the initial value of B.
As a possible implementation manner of this embodiment, the step S4 specifically includes: and calculating the specific position of the sound source according to the time difference from the audio sound source to the audio acquisition equipment.
As a possible implementation manner of this embodiment, the calculating a specific position of the sound source according to a time difference from the audio sound source to the audio acquisition device specifically includes:
setting the coordinates of the audio acquisition equipment of the quaternary array as M1(d/2,0,0), M2(0, d/2,0), M3(-d/2,0,0), M4(0, -d/2,0), M1, M2, M3 and M4 as array elements of the audio acquisition equipment respectively, the coordinates of a sound source as (x, y, z), the distance from the sound source to the origin of coordinates as r, the elevation angle as theta, the azimuth angle as phi and d as the interval of the array elements;
relative to M1, the time delay of sound source arriving at array elements M2, M3 and M4 is tau21，τ31，τ41The sound differences from the sound source to M2, M3 and M4 are d21, d31 and d41 respectively, and the position of the sound source is solved according to the following equation:
solving to obtain the sound source position as follows:
as a possible implementation manner of this embodiment, the audio in the cable tunnel includes a water dropping sound audio, a fire catching sound audio, a cable joint failure sound audio, a tunnel person walking sound audio, a manhole cover opening sound audio, and a vehicle pressure manhole cover sound audio.
Fig. 2 is a block diagram illustrating an audio-based cable tunnel monitoring system in accordance with an exemplary embodiment. As shown in fig. 2, an audio-based cable tunnel monitoring system provided in an embodiment of the present invention includes:
the signal acquisition module is used for acquiring sound in the cable tunnel;
the signal processing module is used for processing the acquired audio signals;
the characteristic extraction module is used for extracting a characteristic value from the acquired audio signal;
the fault identification module is used for identifying faults according to the extracted features;
and the fault positioning module is used for positioning the fault.
The signal acquisition module adopts a plurality of capacitance test microphones (microphones) to acquire audio frequency in the cable tunnel, and the plurality of capacitance test microphones are arranged at each detection point in the cable tunnel and form a microphone matrix. The capacitance test microphone is an ECM888B capacitance test microphone, has a frequency response range of (20-20) kHz, has sensitivity of-43 dB +/-3 dB (1kHz open circuit voltage), and has good output response within a temperature range of (-10-50) DEG C and a relative humidity range of (0-95%).
As a possible implementation manner of this embodiment, the signal processing module is specifically configured to:
carrying out two-stage in-phase proportional amplification processing on the collected sound signals;
filtering the amplified sound signal;
and sampling the sound signal after the filtering processing.
And a two-stage in-phase proportional amplifier is adopted to carry out two-stage in-phase proportional amplification processing on the collected sound signals, and the two-stage amplification factors are respectively controlled by the switching value of the microprocessor. The microprocessor can automatically select the corresponding amplification factor according to the output of the microphone. The initial amplification factor is 1, when the amplitude of the output signal of the microphone is detected to be 20-200mV, the microprocessor adjusts the amplification factor to be 150 times, when the amplitude of the output signal of the microphone is 200mV-1V, the amplification factor is set to be 15 times, and when the amplitude of the output signal is more than 1V, the amplification factor is 1 time. The operational amplifier LF356 has wide gain band and is not easy to vibrate, and the amplification factor is switched by the relay through the I/O of the microprocessor.
The filtering processing is carried out through a filtering circuit, and the filtering circuit realizes the function of a band-pass filter by utilizing a passive first-order RC high-pass filter and an active second-order Butterworth low-pass filter.
The AIC23B is adopted for carrying out the adoption processing, and the AIC23B has the characteristics of high precision, low power consumption, high sampling rate and the like. The AIC23B adopts advanced Sigma-Delta oversampling technology, and the signal-to-noise ratio of ADC and DAC can reach 90dB and 100dB respectively under the condition that the sampling rate is 48kHz, thereby realizing high-fidelity recording and high-quality playing, and supporting IIC and SPI bus mode to access internal registers.
As a possible implementation manner of this embodiment, the feature extraction module extracts a feature value of the audio information from the acquired audio information through Mel frequency cepstrum coefficients that simulate human auditory characteristics. The original signal is processed by the processes of pre-emphasis, framing, windowing, FFT, Mel filter bank, logarithm operation, DCT and the like to obtain the MFCC parameters. The method specifically comprises the following steps:
the audio signal S (n) is preprocessed to obtain a time domain signal X (n) of each audio frame, and the time domain signal is processed by FFT to obtain a linear spectrum X (k).
Performing square operation on the amplitude of the linear spectrum X (k), and obtaining an energy spectrum; the energy spectrum is band-pass filtered. In the frequency domain, a set of Mel-scale triangular filter banks is provided for band-pass filtering the energy spectrum. The Mel-frequency filter bank refers to a plurality of hm (k) band-pass filters arranged in the audio frequency spectrum range.
The transfer function of the band pass filter is:
the logarithmic energy calculation formula output by each filter bank is as follows:
obtaining a voice characteristic parameter MFCC through DCT (discrete cosine transform):
the audio signal extraction process comprises the steps of framing the sound signals of the data set, wherein the frame length is 256, and the frame is 80; 24 digital filters are selected, DCT coefficients are 12-by-24 dimensions, 12 MFCC coefficients obtained after sound signals are subjected to discrete Fourier transform and 12 coefficients of primary difference of the sound signals are calculated, and 24-dimensional features are calculated; the MFCC fusion characteristics take the short-time zero-crossing rate characteristics of signals and the 24-dimensional MFCC characteristics to sum up to 25-dimensional characteristics for experiment.
As a possible implementation manner of this embodiment, the fault identification module is specifically configured to:
the extracted features are a set of random vector sequences O ═ O1o2,…oT；
The vector O is O by vector quantization method1o2,…oTAdjusting the sequence to be an observation sequence needing to be input in a hidden Markov model, and obtaining hidden Markov models of different sound signals for the sequence through a hidden Markov model training algorithm to form a fault model library;
the fault sound signal to be identified is input, and O is obtained by the previous method1o2,…oTThen, the output probability P (O | M) of each hidden Markov model is calculated, and the digital type corresponding to the hidden Markov model with the maximum P (O | M) is the recognition result.
The invention selects the most common Hidden Markov Model (HMM) in the field of audio recognition as a recognition method.
As a possible implementation manner of this embodiment, the hidden markov model training algorithm specifically includes:
a hidden markov model is described by M ═ { S, O, a, B, pi, F }, where S is the set of all states in the HMM, O is the set of observed value symbols that are output, and a ═ { a ═ a }ijI, j is 1,2, …, N }, B is a set of output observation value probabilities, pi is a set of system initial state probabilities, and F is a set of system termination states;
segmenting the audio of the same type, wherein the length of the intercepted audio of different types is 20 frames to 500 frames;
setting initial values of pi and A, B at will, obtaining the hidden state of the audio frequency according to the Viterbi algorithm, and calculating the probability of each observation value of each state according to the estimated hidden state and the observation value to obtain the initial value of B.
When a Hidden Markov Model (HMM) is trained, firstly, the same type of audio is segmented, such as cable joint fault sound, tunnel person walking sound, well lid opening sound, vehicle well lid pressing sound and the like, and the length of the different types of audio is 20 frames to 500 frames. Since the initial values of the observation probability matrix B have a large influence on training the HMM, it is important to calculate the initial values of B. The initial values of a and pi have little effect on training the HMM relative to the importance of B, and according to this feature, the values of a and pi are usually set manually. When estimating the initial value of B, firstly randomly setting initial values of pi and A, B, then obtaining the hidden state of the audio according to the Viterbi algorithm, and then calculating the probability of each observation value of each state according to the estimated hidden state and the observation value, thereby obtaining the initial value of B.
When the initial value of the HMM is determined, it is a cyclic process of reestimating it by the Baum-Welch algorithm. After the process is finished, the Viterbi algorithm calculates the probability of the appearance of the observed value sequence under the reestimation model, and when the probability tends to be stable, the parameter reestimation cycle process is ended. In order to ensure that the trained model is accurate, the number of iterations is selected to be 100.
2) Recognition of Hidden Markov Models (HMMs). After the audio signal with the fault to be recognized is subjected to the preprocessing process, calculating the output probability of the feature sequence of the audio signal to each trained Hidden Markov Model (HMM) by utilizing the trained HMM and adopting a Veterbi algorithm; and selecting the digital type corresponding to the HMM with the maximum output probability as an output result.
As a possible implementation manner of this embodiment, the fault location module is specifically configured to: and calculating the specific position of the sound source according to the time difference from the audio sound source to the audio acquisition equipment.
As a possible implementation manner of this embodiment, the calculating a specific position of the sound source according to a time difference from the audio sound source to the audio acquisition device specifically includes:
sound source localization the sound source is localized according to the plane quaternary cross principle. The sound transmitter group adopts 4 sound transmitter groups, and the specific position of the sound source is calculated according to the time difference from the sound source to each microphone. The sound source processing flow is shown in fig. 3 (the specific steps are not described), a peak value of each microphone is obtained, according to the arrival time difference of the peak value of each microphone, coordinates of four microphones of a quaternary array are assumed to be M1(d/2,0,0), M2(0, d/2,0), M3(-d/2,0,0), M4(0, -d/2,0), coordinates of a sound source are (x, y, z), the distance from the sound source to the coordinate origin is r, the elevation angle is θ, the azimuth angle is Φ, and d is the array element distance. According to the assumption that the time delay of sound source reaching array elements M2, M3 and M4 is tau relative to M121，τ31，τ41The sound differences from the sound source to M2, M3 and M4 are d21, d31 and d41 respectively, and the position of the sound source is solved according to the following equation:
solving to obtain the sound source position as follows:
as a possible implementation manner of this embodiment, the audio in the cable tunnel includes a water dropping sound audio, a fire catching sound audio, a cable joint failure sound audio, a tunnel person walking sound audio, a manhole cover opening sound audio, and a vehicle pressure manhole cover sound audio.
The data transmission adopts a TCP protocol, and a W5500 chip is selected. W5500 supports high-speed standard 4-wire SPI interface to communicate with the host computer, and the SPI rate can reach 80MHz theoretically. The Ethernet data link layer (MAC) and the 10BaseT/100BaseTX Ethernet physical layer (PHY) are integrated inside the network, and the network management system supports automatic negotiation (10/100-Based full duplex/half duplex), power-down mode and network wake-up function. Different from a traditional software protocol stack, 8 independent hardware sockets embedded in the W5500 can carry out 8-way independent communication, the communication efficiency of the 8-way sockets is not affected mutually, and the size of each Socket can be flexibly defined through the receiving/sending cache of 32 Kbytes on the W5500 chip.
An audio-based cable tunnel monitoring device can be formed by an audio-based cable tunnel monitoring system. The equipment adopts a modular design, has a simple structure and is easy to realize, and comprises a front-end acquisition device, a signal transmission device and a background analysis system. Through set up microprocessor at the cable tunnel scene, microprocessor and signal acquisition module, signal processing module and electric capacity test microphone constitute front end collection system, the characteristic draws the module, fault identification module and fault localization module constitute backstage analytic system, backstage analysis software can carry out real-time processing and the analysis of time domain and frequency domain to the sound source signal that the front end equipment gathered, calculate the position of biography origin, and carry out analysis and location to the sound source signal, adopt artificial neural network to train the sound source, realize electric power tunnel sound source automatic identification, the analysis. Data transmission between the front-end acquisition device and the background analysis system adopts a TCP protocol, and the signal transmission device adopts a W5500 chip. W5500 supports high-speed standard 4-wire SPI interface to communicate with the host computer, and the SPI rate can reach 80MHz theoretically. The Ethernet data link layer (MAC) and the 10BaseT/100BaseTX Ethernet physical layer (PHY) are integrated inside the network, and the network management system supports automatic negotiation (10/100-Based full duplex/half duplex), power-down mode and network wake-up function. Different from a traditional software protocol stack, 8 independent hardware sockets embedded in the W5500 can carry out 8-way independent communication, the communication efficiency of the 8-way sockets is not affected mutually, and the size of each Socket can be flexibly defined through the receiving/sending cache of 32 Kbytes on the W5500 chip. The equipment can analyze the sound in the cable tunnel, can analyze and position the sound source, and helps workers to visually analyze the running condition in the tunnel.
The foregoing is only a preferred embodiment of the present invention, and it will be apparent to those skilled in the art that various modifications and improvements can be made without departing from the principle of the present invention, and these modifications and improvements are also considered to be within the scope of the present invention.
Claims (10)
1. A cable tunnel monitoring method based on audio is characterized by comprising the following steps:
s1, collecting audio in the cable tunnel;
s2, processing the collected audio signal;
s3, extracting characteristic values of the collected audio signals;
s4, identifying faults according to the extracted features;
and S5, positioning the fault.
2. The audio-based cable tunnel monitoring method according to claim 1, wherein in step S1, the audio frequency in the cable tunnel is collected by using a plurality of capacitive test microphones, and the plurality of capacitive test microphones are disposed at respective detection points in the cable tunnel and form a microphone matrix.
3. The audio-based cable tunnel monitoring method according to claim 1, wherein the step S2 is specifically:
carrying out two-stage in-phase proportional amplification processing on the collected sound signals;
filtering the amplified sound signal;
and sampling the sound signal after the filtering processing.
4. The audio-based cable tunnel monitoring method according to claim 1, wherein the step S3 is specifically:
the audio signal S (n) is preprocessed to obtain a time domain signal X (n) of each audio frame, and the time domain signal is processed by FFT to obtain a linear spectrum X (k).
Performing square operation on the amplitude of the linear spectrum X (k), and obtaining an energy spectrum; the energy spectrum is band-pass filtered.
The transfer function of the band pass filter is:
the logarithmic energy calculation formula output by each filter bank is as follows:
obtaining a voice characteristic parameter MFCC through DCT (discrete cosine transform):
5. the audio-based cable tunnel monitoring method according to claim 1, wherein the step S4 is specifically:
the extracted features are a set of random vector sequences O ═ O1o2,…oT；
The vector O is O by vector quantization method1o2,…oTAdjusting the sequence to be an observation sequence needing to be input in a hidden Markov model, and obtaining hidden Markov models of different sound signals for the sequence through a hidden Markov model training algorithm to form a fault model library;
the fault sound signal to be identified is input, and O is obtained by the previous method1o2,…oTThen, the output probability P (O | M) of each hidden Markov model is calculated, and the digital type corresponding to the hidden Markov model with the maximum P (O | M) is the recognition result.
6. The audio-based cable tunnel monitoring method according to claim 5, wherein the hidden Markov model training algorithm is specifically:
a hidden markov model is described by M ═ { S, O, a, B, pi, F }, where S is the set of all states in the HMM, O is the set of observed value symbols that are output, and a ═ { a ═ a }ijI, j is 1,2, a.
Segmenting the audio of the same type, wherein the length of the intercepted audio of different types is 20 frames to 500 frames;
setting initial values of pi and A, B at will, obtaining the hidden state of the audio frequency according to the Viterbi algorithm, and calculating the probability of each observation value of each state according to the estimated hidden state and the observation value to obtain the initial value of B.
7. The audio-based cable tunnel monitoring method according to claim 1, wherein the step S4 is specifically: and calculating the specific position of the sound source according to the time difference from the audio sound source to the audio acquisition equipment.
8. The audio-based cable tunnel monitoring method according to claim 7, wherein the specific location of the sound source is calculated according to a time difference from the audio sound source to the audio collecting device, specifically:
setting the coordinates of the audio acquisition equipment of the quaternary array as M1(d/2,0,0), M2(0, d/2,0), M3(-d/2,0,0), M4(0, -d/2,0), M1, M2, M3 and M4 as array elements of the audio acquisition equipment respectively, the coordinates of a sound source as (x, y, z), the distance from the sound source to the origin of coordinates as r, the elevation angle as theta, the azimuth angle as phi and d as the interval of the array elements;
relative to M1, the time delay of sound source arriving at array elements M2, M3 and M4 is tau21，τ31，τ41The sound differences from the sound source to M2, M3 and M4 are d21, d31 and d41 respectively, and the position of the sound source is solved according to the following equation:
solving to obtain the sound source position as follows:
9. the audio-based cable tunnel monitoring method according to any one of claims 1 to 8, wherein the audio in the cable tunnel includes a dripping sound audio, a fire sound audio, a cable joint failure sound audio, a tunnel person walking sound audio, a manhole cover opening sound audio, and a vehicle pressure manhole cover sound audio.
10. An audio-based cable tunnel monitoring system, comprising:
the signal acquisition module is used for acquiring sound in the cable tunnel;
the signal processing module is used for processing the acquired audio signals;
the characteristic extraction module is used for extracting a characteristic value from the acquired audio signal;
the fault identification module is used for identifying faults according to the extracted features;
and the fault positioning module is used for positioning the fault.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010399900.3A CN111540347A (en) | 2020-05-12 | 2020-05-12 | Cable tunnel monitoring method and system based on audio |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010399900.3A CN111540347A (en) | 2020-05-12 | 2020-05-12 | Cable tunnel monitoring method and system based on audio |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111540347A true CN111540347A (en) | 2020-08-14 |
Family
ID=71969434
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010399900.3A Pending CN111540347A (en) | 2020-05-12 | 2020-05-12 | Cable tunnel monitoring method and system based on audio |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111540347A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112466076A (en) * | 2020-12-12 | 2021-03-09 | 江西洪都航空工业集团有限责任公司 | Intelligent well lid anti-theft method |
| CN112505470A (en) * | 2020-09-28 | 2021-03-16 | 华北电力大学 | System and method for fault location in long-distance conveying facility |
| CN112735464A (en) * | 2020-12-21 | 2021-04-30 | 招商局重庆交通科研设计院有限公司 | Tunnel emergency broadcast sound effect information detection method |
| CN112816060A (en) * | 2020-12-29 | 2021-05-18 | 杭州电子科技大学 | Sound and vibration combined monitoring system and analysis method thereof |
| CN113050039A (en) * | 2021-03-10 | 2021-06-29 | 杭州瑞利超声科技有限公司 | Acoustic fluctuation positioning system used in tunnel |
| CN113362077A (en) * | 2021-05-25 | 2021-09-07 | 深圳市中晴云科技有限公司 | Product traceability system and method based on block chain |
| CN113466616A (en) * | 2021-06-22 | 2021-10-01 | 海南电网有限责任公司乐东供电局 | Method and device for quickly positioning cable fault point |
| WO2022105287A1 (en) * | 2020-11-18 | 2022-05-27 | 华北电力大学 | System and method for monitoring abnormal state on the basis of voiceprint diagnosis |
| CN115420798A (en) * | 2022-08-31 | 2022-12-02 | 中国国家铁路集团有限公司 | Tunnel lining cavity detection method and system based on sound pulse signals |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103278801A (en) * | 2013-05-22 | 2013-09-04 | 河南省电力公司周口供电公司 | Noise imaging detection device and detection calculation method for transformer substation |
| CN105244038A (en) * | 2015-09-30 | 2016-01-13 | 金陵科技学院 | Ore dressing equipment fault abnormity audio analyzing and identifying method based on HMM |
| CN105336331A (en) * | 2015-11-21 | 2016-02-17 | 太原理工大学 | Intelligent monitoring method and intelligent monitoring system for abnormal behaviors of pigs on basis of sound |
| CN107424626A (en) * | 2017-08-04 | 2017-12-01 | 上海振华重工电气有限公司 | A kind of gantry crane trolley track sound monitoring and fault finding system |
| CN109009129A (en) * | 2018-08-20 | 2018-12-18 | 南京农业大学 | Sow respiratory disease method for early warning based on acoustic analysis |
| CN109151145A (en) * | 2018-09-07 | 2019-01-04 | 深圳市万普拉斯科技有限公司 | A kind of detection method and device of voice-grade channel |
-
2020
- 2020-05-12 CN CN202010399900.3A patent/CN111540347A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103278801A (en) * | 2013-05-22 | 2013-09-04 | 河南省电力公司周口供电公司 | Noise imaging detection device and detection calculation method for transformer substation |
| CN105244038A (en) * | 2015-09-30 | 2016-01-13 | 金陵科技学院 | Ore dressing equipment fault abnormity audio analyzing and identifying method based on HMM |
| CN105336331A (en) * | 2015-11-21 | 2016-02-17 | 太原理工大学 | Intelligent monitoring method and intelligent monitoring system for abnormal behaviors of pigs on basis of sound |
| CN107424626A (en) * | 2017-08-04 | 2017-12-01 | 上海振华重工电气有限公司 | A kind of gantry crane trolley track sound monitoring and fault finding system |
| CN109009129A (en) * | 2018-08-20 | 2018-12-18 | 南京农业大学 | Sow respiratory disease method for early warning based on acoustic analysis |
| CN109151145A (en) * | 2018-09-07 | 2019-01-04 | 深圳市万普拉斯科技有限公司 | A kind of detection method and device of voice-grade channel |
Non-Patent Citations (2)
| Title |
|---|
| 杜世斌: "基于音频特征的电气设备故障监测算法研究", 《中国优秀硕士学位论文全文数据库 工程科技II辑》 * |
| 陆灏铭等: "基于麦克风阵列的声源定位系统设计", 《传感器与微系统》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112505470A (en) * | 2020-09-28 | 2021-03-16 | 华北电力大学 | System and method for fault location in long-distance conveying facility |
| WO2022062622A1 (en) * | 2020-09-28 | 2022-03-31 | 华北电力大学 | System and method for fault location in long-distance transmission facility |
| WO2022105287A1 (en) * | 2020-11-18 | 2022-05-27 | 华北电力大学 | System and method for monitoring abnormal state on the basis of voiceprint diagnosis |
| CN112466076A (en) * | 2020-12-12 | 2021-03-09 | 江西洪都航空工业集团有限责任公司 | Intelligent well lid anti-theft method |
| CN112735464A (en) * | 2020-12-21 | 2021-04-30 | 招商局重庆交通科研设计院有限公司 | Tunnel emergency broadcast sound effect information detection method |
| CN112816060A (en) * | 2020-12-29 | 2021-05-18 | 杭州电子科技大学 | Sound and vibration combined monitoring system and analysis method thereof |
| CN113050039A (en) * | 2021-03-10 | 2021-06-29 | 杭州瑞利超声科技有限公司 | Acoustic fluctuation positioning system used in tunnel |
| CN113050039B (en) * | 2021-03-10 | 2023-03-07 | 杭州瑞利超声科技有限公司 | Acoustic fluctuation positioning system used in tunnel |
| CN113362077A (en) * | 2021-05-25 | 2021-09-07 | 深圳市中晴云科技有限公司 | Product traceability system and method based on block chain |
| CN113362077B (en) * | 2021-05-25 | 2022-09-23 | 青岛飞熊领鲜数字科技有限公司 | Product traceability system and method based on block chain |
| CN113466616A (en) * | 2021-06-22 | 2021-10-01 | 海南电网有限责任公司乐东供电局 | Method and device for quickly positioning cable fault point |
| CN115420798A (en) * | 2022-08-31 | 2022-12-02 | 中国国家铁路集团有限公司 | Tunnel lining cavity detection method and system based on sound pulse signals |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111540347A (en) | Cable tunnel monitoring method and system based on audio | |
| CN106782504B (en) | Audio recognition method and device | |
| CN106846803B (en) | Traffic event detection device and method based on audio frequency | |
| CN110503972B (en) | Speech enhancement method, system, computer device and storage medium | |
| CN101980336B (en) | Hidden Markov model-based vehicle sound identification method | |
| CN106328151B (en) | ring noise eliminating system and application method thereof | |
| CN101023469A (en) | Digital filtering method, digital filtering equipment | |
| CN103760520B (en) | A kind of single language person sound source DOA method of estimation based on AVS and rarefaction representation | |
| CN109186752B (en) | Underwater acoustic signal acquisition, transmission and detection system based on graphic processor | |
| CN205987367U (en) | Sound pick up equipment based on distributing type microphone | |
| CN106548786A (en) | A kind of detection method and system of voice data | |
| CN111142066A (en) | Direction-of-arrival estimation method, server, and computer-readable storage medium | |
| CN111276150B (en) | Intelligent voice-to-text and simultaneous interpretation system based on microphone array | |
| CN110970020A (en) | Method for extracting effective voice signal by using voiceprint | |
| CN107592600B (en) | Pickup screening method and pickup device based on distributed microphones | |
| CN112420079A (en) | Voice endpoint detection method and device, storage medium and electronic equipment | |
| CN117292713A (en) | Transformer voiceprint monitoring method and system based on end Bian Yun cooperation | |
| CN107547981A (en) | A kind of audio collecting device, supervising device and collection sound method | |
| CN111554319A (en) | Multichannel cardiopulmonary sound abnormity identification system and device based on low-rank tensor learning | |
| CN110706721A (en) | Electric precipitation spark discharge identification method based on BP neural network | |
| CN214796752U (en) | Engineering vehicle identification and positioning device based on voiceprint identification and sound source positioning | |
| CN209402679U (en) | Pickup component and system | |
| CN111627426A (en) | Method and system for eliminating channel difference in voice interaction, electronic equipment and medium | |
| CN108376547A (en) | A kind of array speech collecting system and acquisition method based on Arduino | |
| CN117198312B (en) | Voice interaction processing method for intelligent glasses |
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: 20200814 |
|
| RJ01 | Rejection of invention patent application after publication |