CN108630226A - The artificial intelligence detection method and device of unmanned garage parking O&M safety - Google Patents
The artificial intelligence detection method and device of unmanned garage parking O&M safety Download PDFInfo
- Publication number
- CN108630226A CN108630226A CN201810434317.4A CN201810434317A CN108630226A CN 108630226 A CN108630226 A CN 108630226A CN 201810434317 A CN201810434317 A CN 201810434317A CN 108630226 A CN108630226 A CN 108630226A
- Authority
- CN
- China
- Prior art keywords
- voice signal
- garage parking
- spectrum
- safety
- sound
- 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
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; 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 OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; 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 OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; 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
Abstract
The disclosure discloses a kind of artificial intelligence detection method and device of unmanned garage parking O&M safety, belongs to Artificial technical field of intelligence.The method includes:Obtain the voice signal acquired to lifter apparatus during unmanned garage parking O&M, the voice signal is converted into sound spectrum, staggeredly the sound spectrum is handled using preset three-layer coil lamination and average pond layer, the spectrum signature for extracting the sound spectrum carries out O&M according to the spectrum signature and safely identifies.The artificial intelligence detection method and device of above-mentioned unmanned garage parking O&M safety can carry out O&M safety detection to unmanned garage parking in real time, improve the O&M safety detection efficiency in unmanned garage parking.
Description
Technical field
This disclosure relates to computer application technology, more particularly to a kind of artificial intelligence of unmanned garage parking O&M safety
Detection method and device.
Background technology
It is increased rapidly in automobile quantity and soil more and more rare today, the application in unmanned garage parking are largely delayed
The problem of having solved parking difficulty.The operation principle in unmanned garage parking is vehicle parking on vehicle-carrying plate, makes carrier vehicle by mechanical device
Plate lifting or traversing, vehicle in an orderly manner, is three-dimensionally parked, to realize effective expansion of parking space.
In order to ensure the safe and reliable operation in unmanned garage parking, the detection that O&M safety is carried out to unmanned garage parking is needed.
Traditional method is monitored using personnel's post and security protection video, but when the number of plies in unmanned garage parking is more, the demand to manpower
Excessively high, working strength is excessive and can not be accurately when realize the O&M safety detection in unmanned garage parking, while traditional security protection regards
Frequency monitoring can only the simple prolonged monitor video content of record storage, the O&M safety in unmanned garage parking can not be carried out in real time
Detection needs manually to play back observation video record content to carry out the positioning of safety problem, causes to imitate in the case where there are abnormal conditions
Rate is extremely low.
Invention content
In order to solve that O&M safety detection, and less efficient skill can not be carried out to unmanned garage parking in real time in the related technology
Art problem, present disclose provides a kind of artificial intelligence detection method, device and the terminals of unmanned garage parking O&M safety.
In a first aspect, a kind of artificial intelligence detection method of unmanned garage parking O&M safety is provided, including:
Obtain the voice signal acquired to lifter apparatus during unmanned garage parking O&M;
The voice signal is converted into sound spectrum;
Staggeredly the sound spectrum is handled using preset three-layer coil lamination and average pond layer, extracts the sound
The spectrum signature of sound spectrum;
O&M is carried out according to the spectrum signature to safely identify.
Second aspect, the artificial intelligence detection for providing a kind of unmanned garage parking O&M safety then execute, which is characterized in that
Described device includes:
Sound acquisition module, for obtaining the voice signal acquired to lifter apparatus during unmanned garage parking O&M;
Frequency spectrum conversion module, for the voice signal to be converted to sound spectrum;
O&M security identity module, for being interlocked to the sound audio using preset three-layer coil lamination and average pond layer
Spectrum is handled, and the spectrum signature of the sound spectrum is extracted;
O&M security identity module is safely identified for carrying out O&M according to the spectrum signature.
The third aspect provides a kind of terminal, including memory and processor, and being stored with computer in the memory can
Reading instruction, when the computer-readable instruction is executed by the processor so that the processor execution is described above, and nobody stops
The step of artificial intelligence detection method of garage O&M safety.
Fourth aspect provides a kind of storage medium being stored with computer-readable instruction, the computer-readable instruction
When being executed by one or more processors so that one or more processors execute unmanned garage parking O&M safety described above
The step of artificial intelligence detection method.
The technical scheme provided by this disclosed embodiment can include the following benefits:
When carrying out the O&M safety detection in unmanned garage parking, sound will be converted to the voice signal that lifter apparatus acquires
After sound spectrum, staggeredly the sound spectrum is handled using preset three-layer coil lamination and average pond layer, described in extraction
It carries out O&M after the spectrum signature of sound spectrum to safely identify, the sound to generate in operation by lifter apparatus
Sound can be achieved with being measured in real time the O&M in unmanned garage parking safely, and substantially increase the O&M peace in unmanned garage parking
The accuracy and efficiency that full inspection is surveyed.
It should be understood that above general description and following detailed description is merely exemplary, this can not be limited
It is open.
Description of the drawings
The drawings herein are incorporated into the specification and forms part of this specification, and shows the implementation for meeting the present invention
Example, and in specification together principle for explaining the present invention.
Fig. 1 is a kind of block diagram of device shown according to an exemplary embodiment.
Fig. 2 is a kind of artificial intelligence detection method of unmanned garage parking O&M safety shown according to an exemplary embodiment
Flow chart.
Fig. 3 be the unmanned garage parking O&M safety of Fig. 2 corresponding embodiments artificial intelligence detection method in step S130
A kind of flow chart of specific implementation.
Fig. 4 is according to the flow diagram to being handled sound spectrum shown in an exemplary embodiment.
Fig. 5 be the unmanned garage parking O&M safety of Fig. 3 corresponding embodiments artificial intelligence detection method in step S132
A kind of flow chart of specific implementation.
Fig. 6 is the artificial intelligence detection side according to the unmanned garage parking O&M safety of another kind shown in Fig. 5 corresponding embodiments
The flow chart of method.
Fig. 7 is that a kind of artificial intelligence detection of unmanned garage parking O&M safety shown according to an exemplary embodiment is then held
Capable block diagram.
Fig. 8 is O&M during the artificial intelligence detection of the unmanned garage parking O&M safety shown in Fig. 7 corresponding embodiments then executes
A kind of block diagram of security identity module 140.
Specific implementation mode
Here will explanation be executed to exemplary embodiment in detail, the example is illustrated in the accompanying drawings.Following description is related to
When attached drawing, unless otherwise indicated, the same numbers in different drawings indicate the same or similar elements.Following exemplary embodiment
Described in embodiment do not represent and the consistent all embodiments of the present invention.On the contrary, they be only with it is such as appended
The example of the consistent device and method of some aspects being described in detail in claims, of the invention.
Fig. 1 is a kind of block diagram of device 100 shown according to an exemplary embodiment.Device 100 can be applied to intelligence
The terminals such as mobile phone, computer.
With reference to figure 1, device 100 may include one or more following component:Processing component 101, memory 102, electricity
Source component 103, multimedia component 104, audio component 105, sensor module 107 and communication component 108.
The integrated operation of 101 usual control device 100 of processing component, such as with display, call, data communication, phase
Machine operates and record operates associated operation etc..Processing component 101 may include one or more processors 109 to execute
Instruction, to perform all or part of the steps of the methods described above.In addition, processing component 101 may include one or more modules,
Convenient for the interaction between processing component 101 and other assemblies.For example, processing component 101 may include multi-media module, with convenient
Interaction between multimedia component 104 and processing component 101.
Memory 102 is configured as storing various types of data to support the operation in device 100.These data are shown
Example includes the instruction for any application program or method that operate on the device 100.Memory 102 can be by any kind of
Volatibility or non-volatile memory device or combination thereof are realized, such as static RAM (SRAM), electrically erasable
Except programmable read only memory (EEPROM), Erasable Programmable Read Only Memory EPROM (EPROM), programmable read only memory
(PROM), read-only memory (ROM), magnetic memory, flash memory, disk or CD.There are one also being stored in memory 102
Or multiple modules, the one or more module are configured to be executed by the one or more processors 109, it is following any to complete
All or part of step in shown method.
Power supply module 103 provides electric power for the various assemblies of device 100.Power supply module 103 may include power management system
System, one or more power supplys and other generated with for device 100, management and the associated component of distribution electric power.
Multimedia component 104 is included in the screen of one output interface of offer between described device 100 and user.One
In a little embodiments, screen may include liquid crystal display (LCD) and touch panel (TP).If screen includes touch panel, screen
Curtain may be implemented as touch screen, to receive input signal from the user.Touch panel includes one or more touch sensings
Device is to sense the gesture on touch, slide, and touch panel.The touch sensor can not only sense touch or sliding action
Boundary, but also detect duration and pressure associated with the touch or slide operation.
Audio component 105 is configured as output and/or input audio signal.For example, audio component 105 includes a Mike
Wind (MIC), when device 100 is in operation mode, when such as call model, logging mode and speech recognition mode, microphone by with
It is set to reception external audio signal.The received audio signal can be further stored in memory 102 or via communication set
Part 108 is sent.In some embodiments, audio component 105 further includes a loud speaker, is used for exports audio signal.
Sensor module 107 includes one or more sensors, and the state for providing various aspects for device 100 is commented
Estimate.For example, sensor module 107 can detect the state that opens/closes of device 100, the relative positioning of component, sensor group
Part 107 can be with the position change of 100 1 components of detection device 100 or device and the temperature change of device 100.At some
In embodiment, which can also include Magnetic Sensor, pressure sensor or temperature sensor.
Communication component 108 is configured to facilitate the communication of wired or wireless way between device 100 and other equipment.Device
100 can access the wireless network based on communication standard, such as WiFi, 2G or 3G or combination thereof.In an exemplary implementation
In example, communication component 108 receives broadcast singal or broadcast related information from external broadcasting management system via broadcast channel.
In one exemplary embodiment, the communication component 108 further includes near-field communication (NFC) module, to promote short range communication.Example
Such as, NFC module can be based on radio frequency identification (RFID) technology, Infrared Data Association (IrDA) technology, ultra wide band (UWB) technology,
Bluetooth (BT) technology and other technologies are realized.
In the exemplary embodiment, device 100 can be believed by one or more application application-specific integrated circuit (ASIC), number
Number processor (DSP), digital signal processing appts (DSPD), programmable logic device (PLD), field programmable gate array
(FPGA), controller, microcontroller, microprocessor or other electronic components are realized, for executing following methods.
Wherein, the processor in the terminal of the embodiment is configured as executing:
Obtain the unmanned garage parking image of high-definition camera acquisition;
The unmanned garage parking image is identified using convolutional neural networks algorithm, determines parking library facilities described
Position in unmanned garage parking image;
According to the position of the parking library facilities in the unmanned garage parking image, from the unmanned garage parking figure
The characteristics of image of the parking library facilities is extracted as in;
The existence that safe O&M feature is carried out to the characteristics of image of the parking library facilities judges;
If all safe O&M features exist, it is determined that the garage parking O&M safety.
The concrete mode of processor execution operation will be in related unmanned garage parking O&M safety in terminal in the embodiment
Artificial intelligence detection method embodiment in be described in detail, explanation will be not set forth in detail herein.
Fig. 2 is a kind of artificial intelligence detection method of unmanned garage parking O&M safety shown according to an exemplary embodiment
Flow chart.The artificial intelligence detection method of the unmanned garage parking O&M safety is in the terminals such as smart mobile phone, computer.Such as figure
Shown in 2, the artificial intelligence detection method of the unmanned garage parking O&M safety may comprise steps of.
Step S110 obtains the voice signal acquired to lifter apparatus in unmanned garage parking operational process.
During O&M, lifter apparatus will be in operating state, and lifter apparatus is in running in unmanned garage parking
It will be made a sound when state.
It is understood that the sound letter that unmanned garage parking during normal O&M and during abnormal O&M, is sent out
It is number distinct.
It can obtain sound collection equipment in real time to lifter apparatus to obtain to the voice signal that lifter apparatus acquires
The voice signal of acquisition, can also be obtain sound collection equipment it for the previous period in sound that lifter apparatus is acquired
Signal.
In one exemplary embodiment, the running sound (microphone of lifter apparatus is acquired using directional array microphone
It is no more than 5 meters apart from equipment distance) because the operational range of lifter apparatus is only limited in unmanned garage parking, and nobody stops
Garage is the space of a relative closure, therefore, O&M is being carried out according to the voice signal of lifter apparatus in unmanned garage parking
The voice signal of time in the past need not be associated with when the detection of safety
Voice signal is converted to sound spectrum by step S120.
By carrying out the extraction of audio frequency characteristics from voice signal, voice signal is converted into sound spectrum.
There are many by way of extracting audio frequency characteristics in voice signal, LPC (linear prediction) may be used from sound
Audio frequency characteristics are extracted in signal, can also audio frequency characteristics be extracted from voice signal using deep-neural-network, can also used
Other modes extract audio frequency characteristics from voice signal.
For example, the sample frequency for setting voice signal uses degree of overlapping for o as f, the voice signal of input is x, Fourier
Transformation points are n, and voice signal is converted to sound spectrum (two-dimensional matrix) using following mathematical model:
Y=20 × log10(|fix(nx-o)÷((length(window())-o))|+ε)
Wherein length (window ()) is the Hamming window length of Fourier transform.
Acoustic information is changed into sound spectrum (x, y, p) (x indicates that time, y indicate frequency, the intensity of p representative voices),
Sound spectrum (x, y, p) is similar to two-dimensional image information, is just suitble to the processing of convolutional neural networks.
Step S130 staggeredly handles sound spectrum using preset three-layer coil lamination and average pond layer, extracts
The spectrum signature of sound spectrum.
Convolutional neural networks be LeCun et al. proposed in 1998 be used for Text region, they are referred to as LeNet-5.
Convolution operation is defined based on two-dimensional image structure, defines each low-level image feature in local experiences domain only with a son of input
Collect related, such as topological neighborhood.Topological local limit inside convolutional layer can so that weight matrix is very sparse, so convolution operation
The only part connection of two layers of connection.It is more convenient efficient compared with a dense matrix multiplication is calculated to calculate such matrix multiplication,
In addition the free parameter of more peanut, which can make statistics calculate, more benefits.Possessing in the image of two dimensional topology,
Identical input pattern can occur in different location, and similar value has been more likely to stronger dependence, this mould for data
Type is very important.It may be in the arbitrary translation position of entire figure, so we are in this way to calculate identical local feature
One local feature operator scans in entire figure.It is transformed to a characteristic pattern here it is convolution and input figure.This scanning
It is considered as extracting identical feature in different positions, they are shared weights, more like with biological neural network.Pass through
This design can not only be such that the complexity of model reduces, but also the quantity of network weight is made also greatly to reduce.CNN exploitation rights
The shared mode of value reduces the number of parameters for needing to learn, compared with general forward direction BP algorithm (Error Back Propagation,
Error back propagation) so that training speed and accuracy are greatly improved.CNN, can as a deep learning algorithm
So that the pretreated expense of data reaches minimum.
It is obvious that depth convolutional network needs largely have mark sample to be trained, and also need in training process
Middle progress sample enhancing.And due to huge, one depth convolutional network needs of training of the presence of convolutional coding structure and data volume
Intensive operand, therefore currently the majority depth convolutional network is trained by GPU.
Convolutional neural networks generally use convolution sum converging operation based on operate, but it does not need it is unsupervised by
Layer pre-training strategy.In entire training process, backpropagation is played the role of being very outstanding, swashs additionally by suitable
Final accuracy rate can be improved while training for promotion speed by encouraging function.
Since lifter apparatus running is in garage parking, and unmanned garage parking is the space of a relative closure, therefore, right
Sound spectrum is relatively single made of the voice signal conversion of lifter apparatus acquisition.
Staggeredly sound spectrum is handled by 3 layers of convolutional layer and average pond layer, is being guaranteed accurately to sound
While frequency spectrum carries out O&M safety identification, identification process is enormously simplified, improves and sound spectrum progress O&M is known safely
Other efficiency.
Also, since maximum pond layer is typically to texture and the side for extracting objects in images or human body as far as possible
Edge information, and average pondizationization is more focused on the background information of extraction content than maximum pond layer, and reduce the knowledge of adjacent spectra
Other error replaces the maximum pond layer in traditional convolution neural network algorithm by using average pond layer, more ensure that
The identification accuracy in the safe identification process of O&M is carried out to sound spectrum using convolutional neural networks algorithm.
For example, as shown in figure 3, input sound spectrum be pretreated the data matrix for 112*112*1, then lead to successively
Cross the convolutional layer, average pond layer, the convolutional layer of 7*7*128, averagely pond layer, the convolutional layer of 5*5*256, average pond of 9*9*48
Change layer to be handled, output result carries out Classification and Identification by a softmax for sound spectrum.
By the staggeredly processing of convolutional layer and average pond layer, spectrum signature is extracted from sound spectrum.
Step S140 carries out O&M according to spectrum signature and safely identifies.
Lifter apparatus will be adopted using method as described above when carrying out the O&M safety detection in unmanned garage parking
After the voice signal of collection is converted to sound spectrum, using preset three-layer coil lamination with average pond layer staggeredly to the sound audio
Spectrum is handled, and progress O&M safely identifies after extracting the spectrum signature of the sound spectrum, to pass through lifter apparatus
The sound generated in operation can be achieved with being measured in real time the O&M in unmanned garage parking safely, and greatly improve
The accuracy and efficiency of the O&M safety detection in unmanned garage parking.
Optionally, the artificial intelligence detection of the unmanned garage parking O&M safety shown in exemplary embodiment is corresponded to according to fig. 2
Method, before step S120, the artificial intelligence detection method of the unmanned garage parking O&M safety can also include the following steps:
Voice signal is filtered.
Optionally, in the operation of lifter apparatus, the other equipment in unmanned garage parking also will more or less
It makes a sound, there are other noises in the voice signal to make acquisition.
Therefore, O&M safety detection is carried out to unmanned garage parking by the voice signal of lifter apparatus to further increase
Accuracy, the sound sent out to the other equipment in unmanned garage parking is filtered.
It is filtered by the voice signal to acquisition, rejects the sound that the other equipment in unmanned garage parking is sent out.
There are many modes being filtered to the voice signal of acquisition, and sound of the Wiener filtering mode to acquisition may be used
Signal is filtered, can also LMS (Least Mean Square, least mean-square error) algorithms to the voice signal of acquisition into
Row filtering can also use other modes to be filtered the voice signal to acquisition, herein without describing one by one.
Using method as described above, before the voice signal of acquisition is converted to sound spectrum, to voice signal into
Row filtering, carries out voice signal the filtration treatment of noise, other noises in voice signal is avoided to safely identify O&M
It interferes, further improves the accuracy to the O&M safety detection in unmanned garage parking by the voice signal of acquisition.
Optionally, Fig. 4 is the description for corresponding to the details to step S140 shown in exemplary embodiment according to fig. 2.Such as Fig. 4
Shown, step S140 may comprise steps of.
Step S141 carries out spectrum signature by preset full Connection Neural Network the calculating of safe O&M probability.
Safe O&M probability is the possibility degree of lifter apparatus spectrum signature when spectrum signature meets safe O&M.
Full Connection Neural Network is preformed.
Spectrum signature is carried out by full Connection Neural Network to return classification processing, calculates the sound belonging to the spectrum signature
Signal meets the safe O&M probability of safe O&M.
Step S142 carries out unmanned garage parking according to safe O&M probability the judgement of O&M safety.
In one exemplary embodiment, by presetting a safe O&M probability critical value, by safe O&M probability and peace
Dimension probability critical value for the national games is compared, when safe O&M probability reaches safe O&M probability critical value, it is determined that nobody stops
Garage O&M safety;When safe O&M probability is not up to safe O&M probability critical value, it is determined that unmanned garage parking O&M is not
Safety.
For example, being 99.58% in the safe O&M probability being calculated, to critical according to the safe O&M probability of setting
Value (assuming that safe O&M probability critical value is 80%) determines unmanned garage parking O&M safety.
Pass through preset full connection nerve net after the spectrum signature of extraction sound spectrum using method as described above
Network carries out spectrum signature the calculating of safe O&M probability, and then realizes the judgement that O&M safety is carried out to unmanned garage parking, into
One step ensure that the accuracy for carrying out parking library facilities detection identification.
Optionally, Fig. 5 is the people that the unmanned garage parking O&M safety of another kind shown in exemplary embodiment is corresponded to according to Fig. 4
Work intelligent detecting method, before step S130, the artificial intelligence detection method of the unmanned garage parking O&M safety can also wrap
Include following steps.
Step S210 acquires the sound of lifter apparatus respectively during unmanned garage parking is in normal O&M and abnormal O&M
Sound signal obtains positive sample voice signal collection and negative sample voice signal collection.
Positive sample voice signal collection is the sound acquired to lifter apparatus during safe O&M to unmanned garage parking
Signal.
Negative sample voice signal collection is the sound acquired to lifter apparatus during abnormal O&M to unmanned garage parking
Signal.
For example, when carrying out the formation of positive sample voice signal collection, transported by confirming that unmanned garage parking is in safety in advance
During dimension, voice signal when carrying out normal operation in 240 hours to lifter apparatus is acquired;Carrying out negative sample sound
When the formation of signal collection, during the abnormal operation of lifter apparatus or the artificial sound that abnormal sound and abnormal sound is added
Signal is acquired.
Step S230, using convolutional neural networks algorithm respectively to positive sample voice signal collection and negative sample voice signal collection
It is iterated training, forms full Connection Neural Network.
Normally, the number of plies of full Connection Neural Network will not be limited, theoretically the number of plies of full Connection Neural Network
It is more, it is higher in the accuracy for carrying out safe O&M detection by full Connection Neural Network, but also get over consumption calculations resource.
So needing to do balance choice, the number of plies of full Connection Neural Network is enough all right.
Since lifter apparatus running is in garage parking, and unmanned garage parking is the space of a relative closure, therefore, right
Sound spectrum is relatively single made of the voice signal conversion of lifter apparatus acquisition.Namely here to lifter apparatus acquisition
Voice signal is not very complicated, and spectrum signature obtained from extraction is not too many, so using two layers of full connection nerve here
Last two layers in network, such as convolutional neural networks shown in Fig. 4.
Optionally, before step S230, the artificial intelligence detection method of the unmanned garage parking O&M safety can also wrap
Include step S220.
Step S220 respectively expands positive sample voice signal collection and negative sample voice signal collection by microphone response
Exhibition is handled.
In practical applications, sound collection equipment is typically diversified, and the letter that different type microphone is brought
Road distortion is also different.The type of microphone includes:Capacitance microphone, silk ribbon microphone and dynamic microphones etc..Different channels
With different frequency responses, therefore channel distortion can be introduced.This channel distortion can be quite big, especially cheap or low-quality
The microphone of amount.
Various in view of the microphone type actually used, the microphone may each to use acquires individual positive sample sound
Sound signal collection and negative sample voice signal collection, which are trained, will greatly increase the acquisition cost of voice signal.Therefore, by using
The method of microphone impulse response simulates the acquisition for generating different microphones to voice signal, this will significantly reduce acquisition at
Sheet and period.
For example, y=s*i, wherein y, s respectively represent the voice signal of acquisition, filtered voice signal, and i indicates Mike
Wind impulse response,
The voice signal of acquisition is handled by using multiple disclosed microphone impulse responses, simulates different wheats
Gram wind carries out the acquisition of voice signal, to realize the extension of voice signal.
Using method as described above, by acquiring in advance during unmanned garage parking is in normal O&M and abnormal O&M
The voice signal of lifter apparatus obtains positive sample voice signal collection and negative sample voice signal collection, using convolutional neural networks
Algorithm is iterated training to positive sample voice signal collection and negative sample voice signal collection respectively, forms full Connection Neural Network,
And then O&M is carried out to the sound characteristic extracted from sound spectrum by full Connection Neural Network and is safely identified, it ensure that identification
Accuracy.
Following is embodiment of the present disclosure, can be used for executing the artificial intelligence of this above-mentioned unmanned garage parking O&M safety
Detection method embodiment.For those undisclosed details in the apparatus embodiments, the unmanned garage parking O&M of the disclosure is please referred to
The artificial intelligence detection method embodiment of safety.
Fig. 6 is that a kind of artificial intelligence detection of unmanned garage parking O&M safety shown according to an exemplary embodiment is then held
Capable block diagram, the device include but not limited to:Sound acquisition module 110, frequency spectrum conversion module 120, spectrum signature extraction module
130 and O&M security identity module 140,.
Sound acquisition module 110, for obtaining the sound letter acquired to lifter apparatus during unmanned garage parking O&M
Number;
Frequency spectrum conversion module 120, for the voice signal to be converted to sound spectrum;
Spectrum signature extraction module 130, for being interlocked to the sound using preset three-layer coil lamination and average pond layer
Sound spectrum is handled, and the spectrum signature of the sound spectrum is extracted;
O&M security identity module 140 is safely identified for carrying out O&M according to the spectrum signature.
The function of modules and the realization process of effect specifically refer to above-mentioned unmanned garage parking O&M peace in above-mentioned apparatus
The realization process of step is corresponded in full artificial intelligence detection method, details are not described herein.
Optionally, the artificial intelligence detection of unmanned garage parking O&M safety illustrated in fig. 6, which then executes, further includes but unlimited
In:Filter module.
Filter module, for being filtered to the voice signal.
Optionally, as shown in fig. 7, the O&M security identity module 140 shown in Fig. 6 includes but not limited to:Safe O&M is general
Rate computing unit 141 and O&M analysis unit 142.
Safe O&M probability calculation unit 141, for by preset full Connection Neural Network to the spectrum signature into
The calculating of the safe O&M probability of row;
O&M analysis unit 142, for carrying out O&M to the unmanned garage parking according to the safe O&M probability
The judgement of safety.
Optionally, as shown in figure 8, the artificial intelligence detection of unmanned garage parking O&M safety illustrated in fig. 6 then executes also
Including but not limited to:Sample audio signal collection generation module 210 and repetitive exercise module 230.
Sample audio signal collection generation module 210, for during unmanned garage parking is in normal O&M and abnormal O&M
The voice signal for acquiring lifter apparatus respectively, obtains positive sample voice signal collection and negative sample voice signal collection;
Repetitive exercise module 230, for using convolutional neural networks algorithm respectively to the positive sample voice signal collection and
The negative sample voice signal collection is iterated training, forms the full Connection Neural Network.
Optionally, the artificial intelligence detection of unmanned garage parking O&M safety illustrated in fig. 8, which then executes, further includes but unlimited
In:Extension process module.
Extension process module, for being responded respectively to the positive sample voice signal collection and the negative sample by microphone
Voice signal collection is extended processing.
Optionally, the present invention also provides a kind of terminal, execute as the above exemplary embodiments it is any shown in unmanned stop
The all or part of step of the artificial intelligence detection method of library O&M safety.Terminal includes:
Processor;And
The memory being connect with the processor communication;Wherein,
The memory is stored with readable instruction, and the readable instruction is realized when being executed by the processor as above-mentioned
Method described in either exemplary embodiment.
Processor executes the concrete mode of operation in the related unmanned garage parking O&M in terminal in the embodiment
Detailed description is performed in the embodiment of the artificial intelligence detection method of safety, explanation will be not set forth in detail herein.
In the exemplary embodiment, a kind of storage medium is additionally provided, which is that computer readable storage is situated between
Matter, such as can be the provisional and non-transitory computer readable storage medium for including instruction.The storage medium for example wraps
The memory 102 of instruction is included, above-metioned instruction can be executed by the processor 109 of terminal 100 to complete above-mentioned unmanned garage parking O&M
The artificial intelligence detection method of safety.
It should be understood that the invention is not limited in the precision architectures for being described above and being shown in the accompanying drawings, and
And various modifications and change can be being executed without departing from the scope.The scope of the present invention is limited only by the attached claims.
Claims (10)
1. a kind of artificial intelligence detection method of unmanned garage parking O&M safety, which is characterized in that the method includes:
Obtain the voice signal acquired to lifter apparatus during unmanned garage parking O&M;
The voice signal is converted into sound spectrum;
Staggeredly the sound spectrum is handled using preset three-layer coil lamination and average pond layer, extracts the sound audio
The spectrum signature of spectrum;
O&M is carried out according to the spectrum signature to safely identify.
2. according to the method described in claim 1, it is characterized in that, the step that the voice signal is converted to sound spectrum
Before rapid, the method further includes:
The voice signal is filtered.
3. according to the method described in claim 1, it is characterized in that, the knowledge for carrying out O&M safety according to the spectrum signature
Other step includes:
The calculating of safe O&M probability is carried out to the spectrum signature by preset full Connection Neural Network;
The judgement of O&M safety is carried out to the unmanned garage parking according to the safe O&M probability.
4. according to the method described in claim 3, it is characterized in that, it is described by preset full Connection Neural Network to the frequency
Before spectrum signature carries out the step of calculating of safe O&M probability, the method further includes:
The voice signal for acquiring lifter apparatus respectively during unmanned garage parking is in normal O&M and abnormal O&M, obtains just
Sample audio signal collection and negative sample voice signal collection;
The positive sample voice signal collection and the negative sample voice signal collection are carried out respectively using convolutional neural networks algorithm
Repetitive exercise forms the full Connection Neural Network.
5. according to the method described in claim 4, it is characterized in that, described use convolutional neural networks algorithm respectively to positive sample
Voice signal collection and negative sample voice signal collection are iterated training, before the step of forming the full Connection Neural Network, institute
The method of stating further includes:
Place is extended to the positive sample voice signal collection and the negative sample voice signal collection respectively by microphone response
Reason.
6. a kind of artificial intelligence detection device of unmanned garage parking O&M safety, which is characterized in that described device includes:
Sound acquisition module, for obtaining the voice signal acquired to lifter apparatus during unmanned garage parking O&M;
Frequency spectrum conversion module, for the voice signal to be converted to sound spectrum;
O&M security identity module, for using preset three-layer coil lamination and average pond layer interlock to the sound spectrum into
Row processing, extracts the spectrum signature of the sound spectrum;
O&M security identity module is safely identified for carrying out O&M according to the spectrum signature.
7. device according to claim 6, which is characterized in that described device further includes:
Filter module, for being filtered to the voice signal.
8. device according to claim 6, which is characterized in that the O&M security identity module includes:
Safe O&M probability calculation unit, for carrying out safe fortune to the spectrum signature by preset full Connection Neural Network
Tie up the calculating of probability;
O&M analysis unit is sentenced for carrying out O&M safety to the unmanned garage parking according to the safe O&M probability
It is disconnected.
9. device according to claim 8, which is characterized in that described device further includes:
Sample audio signal collection generation module, for being acquired respectively during unmanned garage parking is in normal O&M and abnormal O&M
The voice signal of lifter apparatus obtains positive sample voice signal collection and negative sample voice signal collection;
Repetitive exercise module, for using convolutional neural networks algorithm respectively to the positive sample voice signal collection and the negative sample
This voice signal collection is iterated training, forms the full Connection Neural Network.
10. device according to claim 9, which is characterized in that described device further includes:
Extension process module, for being responded respectively to the positive sample voice signal collection and the negative sample sound by microphone
Signal collection is extended processing.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810434317.4A CN108630226A (en) | 2018-05-08 | 2018-05-08 | The artificial intelligence detection method and device of unmanned garage parking O&M safety |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810434317.4A CN108630226A (en) | 2018-05-08 | 2018-05-08 | The artificial intelligence detection method and device of unmanned garage parking O&M safety |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108630226A true CN108630226A (en) | 2018-10-09 |
Family
ID=63696075
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810434317.4A Pending CN108630226A (en) | 2018-05-08 | 2018-05-08 | The artificial intelligence detection method and device of unmanned garage parking O&M safety |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108630226A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109473120A (en) * | 2018-11-14 | 2019-03-15 | 辽宁工程技术大学 | A kind of abnormal sound signal recognition method based on convolutional neural networks |
CN110763685A (en) * | 2019-10-22 | 2020-02-07 | 陕西源杰半导体技术有限公司 | Artificial intelligent detection method and device for DFB semiconductor laser chip surface defects |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1993285A (en) * | 2004-08-18 | 2007-07-04 | 东芝电梯株式会社 | Elevator troubleshooting apparatus |
CN102556792A (en) * | 2012-03-13 | 2012-07-11 | 杭州市特种设备检测院 | Online analysis meter and online analytical method for operation performance of elevator |
CN102718100A (en) * | 2012-06-13 | 2012-10-10 | 中山市卓梅尼控制技术有限公司 | Early warning system for elevator fault and early warning method for elevator fault |
CN105424395A (en) * | 2015-12-15 | 2016-03-23 | 珠海格力电器股份有限公司 | Method and device for determining equipment fault |
US20160125892A1 (en) * | 2014-10-31 | 2016-05-05 | At&T Intellectual Property I, L.P. | Acoustic Enhancement |
US20160125894A1 (en) * | 2014-10-31 | 2016-05-05 | At&T Intellectual Property I, L.P. | Self-Organized Acoustic Signal Cancellation Over a Network |
CN106710589A (en) * | 2016-12-28 | 2017-05-24 | 百度在线网络技术(北京)有限公司 | Artificial intelligence-based speech feature extraction method and device |
CN106847302A (en) * | 2017-02-17 | 2017-06-13 | 大连理工大学 | Single channel mixing voice time-domain seperation method based on convolutional neural networks |
CN107221320A (en) * | 2017-05-19 | 2017-09-29 | 百度在线网络技术(北京)有限公司 | Train method, device, equipment and the computer-readable storage medium of acoustic feature extraction model |
CN107522053A (en) * | 2017-07-11 | 2017-12-29 | 浙江新再灵科技股份有限公司 | A kind of elevator platform detecting system and method based on audio analysis |
-
2018
- 2018-05-08 CN CN201810434317.4A patent/CN108630226A/en active Pending
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1993285A (en) * | 2004-08-18 | 2007-07-04 | 东芝电梯株式会社 | Elevator troubleshooting apparatus |
CN102556792A (en) * | 2012-03-13 | 2012-07-11 | 杭州市特种设备检测院 | Online analysis meter and online analytical method for operation performance of elevator |
CN102718100A (en) * | 2012-06-13 | 2012-10-10 | 中山市卓梅尼控制技术有限公司 | Early warning system for elevator fault and early warning method for elevator fault |
US20160125892A1 (en) * | 2014-10-31 | 2016-05-05 | At&T Intellectual Property I, L.P. | Acoustic Enhancement |
US20160125894A1 (en) * | 2014-10-31 | 2016-05-05 | At&T Intellectual Property I, L.P. | Self-Organized Acoustic Signal Cancellation Over a Network |
CN105424395A (en) * | 2015-12-15 | 2016-03-23 | 珠海格力电器股份有限公司 | Method and device for determining equipment fault |
CN106710589A (en) * | 2016-12-28 | 2017-05-24 | 百度在线网络技术(北京)有限公司 | Artificial intelligence-based speech feature extraction method and device |
CN106847302A (en) * | 2017-02-17 | 2017-06-13 | 大连理工大学 | Single channel mixing voice time-domain seperation method based on convolutional neural networks |
CN107221320A (en) * | 2017-05-19 | 2017-09-29 | 百度在线网络技术(北京)有限公司 | Train method, device, equipment and the computer-readable storage medium of acoustic feature extraction model |
CN107522053A (en) * | 2017-07-11 | 2017-12-29 | 浙江新再灵科技股份有限公司 | A kind of elevator platform detecting system and method based on audio analysis |
Non-Patent Citations (1)
Title |
---|
王乃峰: "基于深层神经网络的音频特征提取及场景识别研究", 《中国硕士学位论文全文数据库,信息科技辑》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109473120A (en) * | 2018-11-14 | 2019-03-15 | 辽宁工程技术大学 | A kind of abnormal sound signal recognition method based on convolutional neural networks |
CN110763685A (en) * | 2019-10-22 | 2020-02-07 | 陕西源杰半导体技术有限公司 | Artificial intelligent detection method and device for DFB semiconductor laser chip surface defects |
CN110763685B (en) * | 2019-10-22 | 2020-12-08 | 陕西源杰半导体技术有限公司 | Artificial intelligent detection method and device for DFB semiconductor laser chip surface defects |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109841226A (en) | A kind of single channel real-time noise-reducing method based on convolution recurrent neural network | |
CN109473120A (en) | A kind of abnormal sound signal recognition method based on convolutional neural networks | |
CN107609598A (en) | Image authentication model training method, device and readable storage medium storing program for executing | |
CN106940794A (en) | A yard adjoint system is detectd in a kind of target collection | |
CN106710599A (en) | Particular sound source detection method and particular sound source detection system based on deep neural network | |
CN106599866A (en) | Multidimensional user identity identification method | |
CN109473119B (en) | Acoustic target event monitoring method | |
CN112735473B (en) | Method and system for identifying unmanned aerial vehicle based on voice | |
CN109033780B (en) | A kind of edge calculations access authentication method based on wavelet transformation and neural network | |
CN110175526A (en) | Dog Emotion identification model training method, device, computer equipment and storage medium | |
CN106778559A (en) | The method and device of In vivo detection | |
CN104463194A (en) | Driver-vehicle classification method and device | |
CN114220458B (en) | Voice recognition method and device based on array hydrophone | |
CN108630226A (en) | The artificial intelligence detection method and device of unmanned garage parking O&M safety | |
CN103994820B (en) | A kind of moving target recognition methods based on micropore diameter microphone array | |
CN110348434A (en) | Camera source discrimination method, system, storage medium and calculating equipment | |
CN114201989A (en) | Alternating current motor bearing fault diagnosis method adopting convolutional neural network and bidirectional long-time and short-time memory network | |
CN110633689B (en) | Face recognition model based on semi-supervised attention network | |
Yang et al. | Research on subway pedestrian detection algorithms based on SSD model | |
CN105138886A (en) | Robot biometric identification system | |
CN113689382B (en) | Tumor postoperative survival prediction method and system based on medical images and pathological images | |
Tang et al. | Transound: Hyper-head attention transformer for birds sound recognition | |
CN111667002B (en) | Currency identification method, identification device and electronic equipment | |
CN109558803A (en) | SAR target discrimination method based on convolutional neural networks Yu NP criterion | |
CN111582382B (en) | State identification method and device and electronic equipment |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181009 |