CN109406445A - Liquid identification method, feature extracting method, Liquid identification device and storage device - Google Patents
Liquid identification method, feature extracting method, Liquid identification device and storage device Download PDFInfo
- Publication number
- CN109406445A CN109406445A CN201811150314.4A CN201811150314A CN109406445A CN 109406445 A CN109406445 A CN 109406445A CN 201811150314 A CN201811150314 A CN 201811150314A CN 109406445 A CN109406445 A CN 109406445A
- Authority
- CN
- China
- Prior art keywords
- time
- domain signal
- signal map
- liquid
- map
- 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
- 239000007788 liquid Substances 0.000 title claims abstract description 108
- 238000000034 method Methods 0.000 title claims abstract description 50
- 239000013598 vector Substances 0.000 claims abstract description 53
- 238000012545 processing Methods 0.000 claims abstract description 24
- 238000012360 testing method Methods 0.000 claims abstract description 24
- 238000001228 spectrum Methods 0.000 claims description 39
- 238000012706 support-vector machine Methods 0.000 claims description 10
- 239000011159 matrix material Substances 0.000 claims description 8
- 238000013528 artificial neural network Methods 0.000 claims description 5
- 238000000513 principal component analysis Methods 0.000 claims description 5
- 238000011426 transformation method Methods 0.000 claims description 3
- 238000004458 analytical method Methods 0.000 claims description 2
- 239000012530 fluid Substances 0.000 claims description 2
- 239000000470 constituent Substances 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 14
- 238000005516 engineering process Methods 0.000 description 7
- 235000013361 beverage Nutrition 0.000 description 6
- 239000000284 extract Substances 0.000 description 5
- 238000001514 detection method Methods 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 238000002790 cross-validation Methods 0.000 description 2
- 235000015092 herbal tea Nutrition 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 238000001328 terahertz time-domain spectroscopy Methods 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 238000001069 Raman spectroscopy Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000001427 coherent effect Effects 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- OEBRKCOSUFCWJD-UHFFFAOYSA-N dichlorvos Chemical compound COP(=O)(OC)OC=C(Cl)Cl OEBRKCOSUFCWJD-UHFFFAOYSA-N 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 239000008157 edible vegetable oil Substances 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 239000008267 milk Substances 0.000 description 1
- 210000004080 milk Anatomy 0.000 description 1
- 235000013336 milk Nutrition 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- FONRCZUZCHXWBD-VGOFMYFVSA-N n-[3-[[5-chloro-4-(1h-indol-3-yl)pyrimidin-2-yl]amino]phenyl]-3-[[(e)-4-(dimethylamino)but-2-enoyl]amino]benzamide Chemical compound CN(C)C\C=C\C(=O)NC1=CC=CC(C(=O)NC=2C=C(NC=3N=C(C(Cl)=CN=3)C=3C4=CC=CC=C4NC=3)C=CC=2)=C1 FONRCZUZCHXWBD-VGOFMYFVSA-N 0.000 description 1
- OBJNFLYHUXWUPF-IZZDOVSWSA-N n-[3-[[5-chloro-4-(1h-indol-3-yl)pyrimidin-2-yl]amino]phenyl]-4-[[(e)-4-(dimethylamino)but-2-enoyl]amino]benzamide Chemical compound C1=CC(NC(=O)/C=C/CN(C)C)=CC=C1C(=O)NC1=CC=CC(NC=2N=C(C(Cl)=CN=2)C=2C3=CC=CC=C3NC=2)=C1 OBJNFLYHUXWUPF-IZZDOVSWSA-N 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 239000003208 petroleum Substances 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000002310 reflectometry Methods 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000000411 transmission spectrum Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/3581—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using far infrared light; using Terahertz radiation
- G01N21/3586—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using far infrared light; using Terahertz radiation by Terahertz time domain spectroscopy [THz-TDS]
Landscapes
- Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Health & Medical Sciences (AREA)
- Toxicology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The application provides a kind of Liquid identification method, feature extracting method, Liquid identification device and storage device, which includes the first time-domain signal map of the testing liquid that acquisition is placed in a reservoir；Pretreatment is carried out to the first time-domain signal map and obtains the second time-domain signal map；Characteristic processing is carried out to obtain the main feature vector of the second time-domain signal map to the second time-domain signal map；Main feature vector is inputted default classifier to classify；Default classifier classification results are obtained, to obtain the type of liquid to be detected.Through the above way, the application can be by different using reflecting properties of the different liquids to Terahertz, the first time-domain signal map that it is obtained is also different, to extract the main feature vector of the first time-domain signal map corresponding to testing liquid, and default classifier is provided to carry out identification calculating to main feature vector, it, being capable of reduced cost significantly to obtain the type of testing liquid, and recognition speed is fast, recognition accuracy is high.
Description
Technical field
This application involves Liquid identification fields, know more particularly to a kind of Liquid identification method, feature extracting method, liquid
Other device and storage device.
Background technique
In order to eliminate flammable and combustible liquids security risk caused by the public places such as airport, subway station, railway station and
Meet quick, accurate, lossless testing requirements, there is an urgent need to more advanced technologies for public security organ to carry out safety check, existing
Non-damage detection technique mainly has Raman spectroscopy, x-ray method and microwave method, but the equipment of its detection is huge, and price
It is expensive.
Terahertz time-domain spectroscopy (Terahertz Time-Domain Spectroscopy, THz-TDS) technology is one
Relatively new coherent detection technology has been widely used for the fields such as aerospace, biomedicine, substance detection, Homeland Security.
Common THz-TDS system can be divided into two classes: transmission-type and reflective.Each substance is measured using transmission-type THz-TDS technology
Transmission spectrum, extract optical parameter, the work for material identification is relatively more, and technology and algorithm are also more mature.So
And transmission-type configuration needs and two end in contact of sample, being especially to liquid has the polar liquid absorbed by force very much in terahertz wave band
Body, transmission-type THz-TDS technology encounter very big bottleneck on practicability and versatility.In contrast, it is reflective configuration by
In THz wave only on container inner wall interface and liquid effects, superiority has been shown on the problem of liquid absorbs by force.
Summary of the invention
The application provide it is a kind of based on this application involves Liquid identification field, more particularly to a kind of Liquid identification method,
Feature extracting method, Liquid identification device and storage device.The problem of being able to solve Liquid identification higher cost in the prior art.
In order to solve the above technical problems, the technical solution that the application uses is: providing a kind of Liquid identification method, adopt
Collection places the first time-domain signal map of testing liquid in a reservoir；Pretreatment is carried out to the first time-domain signal map to obtain
Obtain the second time-domain signal map；Characteristic processing is carried out to obtain the second time-domain signal figure to the second time-domain signal map
The main feature vector of spectrum；The main feature vector is inputted default classifier to classify；Obtain the default classifier classification
As a result, to obtain the type of the liquid to be detected.
In order to solve the above technical problems, another technical solution that the application uses is: the feature for providing a kind of liquid mentions
Method is taken, the first time-domain signal map including acquiring liquid；Pretreatment is carried out to the first time-domain signal map and obtains the
Two time-domain signal maps；Characteristic processing is carried out to obtain the second time-domain signal map to the second time-domain signal map
Feature vector；Wherein, described eigenvector is above-mentioned main feature vector.
In order to solve the above technical problems, another technical solution that the application uses is: a kind of Liquid identification device is provided,
The device includes processor, memory and Terahertz transceiver, and the processor connects the memory and the Terahertz
Transceiver；Wherein, the Terahertz transceiver is used to acquire the first time-domain signal map for placing testing liquid in a reservoir；
The processor is used to carry out the first time-domain signal map pretreatment to obtain the second time-domain signal map；To described second
Time-domain signal map carries out characteristic processing to obtain the main feature vector of the second time-domain signal map；By the main feature to
Amount inputs default classifier and classifies；The default classifier classification results are obtained, to obtain the class of the liquid to be detected
Type.
In order to solve the above technical problems, another technical solution that the application uses is: a kind of storage device is provided, including
It is able to achieve the program file of the above method.
The beneficial effect of the application is: being different from the prior art, the acquisition prepare liquid that the application passes through acquisition testing liquid
First time-domain signal map of body then carries out pretreatment with characteristic processing to obtain its main spy to the first time-domain signal map
Vector is levied, and main feature vector is input to default classifier and is classified, to obtain the type of testing liquid.Pass through utilization
Different liquids it is different to the reflecting properties of Terahertz, the first obtained time-domain signal map is also different, thus extract to
The main feature vector of the first time-domain signal map corresponding to liquid is surveyed, and default classifier is provided to carry out main feature vector
Identification calculates, to obtain the type of testing liquid, can reduced cost significantly, and recognition speed is high, and recognition accuracy is high.
Detailed description of the invention
Fig. 1 is the flow diagram of the first embodiment of the application Liquid identification method；
Fig. 2 is the schematic diagram that the time domain map of testing liquid is acquired in the application Liquid identification method；
Fig. 3 is the schematic diagram of the application Liquid identification method time domain map collected；
Fig. 4 is the flow diagram of the second embodiment of the application Liquid identification method；
Fig. 5 is the flow diagram of the third embodiment of the application Liquid identification method；
Fig. 6 is the flow diagram of the first embodiment of the application fluid characteristics extracting method；
Fig. 7 is the structural schematic block diagram of one embodiment of the application Liquid identification device；
Fig. 8 is the structural schematic diagram of one embodiment of the application storage device.
Embodiment
Below in conjunction with the attached drawing in the embodiment of the present application, technical solutions in the embodiments of the present application carries out clear, complete
Site preparation description, it is clear that described embodiment is only a part of the embodiment of the application, instead of all the embodiments.It is based on
Embodiment in the application, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall in the protection scope of this application.
In addition, being somebody's turn to do " first ", " second " etc. if relating to the description of " first ", " second " etc. in the embodiment of the present application
Description be used for description purposes only, be not understood to indicate or imply its relative importance or implicitly indicate indicated skill
The quantity of art feature." first " is defined as a result, the feature of " second " can explicitly or implicitly include at least one spy
Sign.It in addition, the technical solution between each embodiment can be combined with each other, but must be with those of ordinary skill in the art's energy
It is enough realize based on, will be understood that the knot of this technical solution when conflicting or cannot achieve when occurs in the combination of technical solution
Conjunction is not present, also not the present invention claims protection scope within.
Referring to Fig. 1, Fig. 1 is a kind of first embodiment flow diagram of Liquid identification method of the application, it is specific to wrap
Include following steps:
The first time-domain signal map of testing liquid in a reservoir is placed in S11, acquisition.
Currently, liquid container on the market mainly has canister, plastic containers or paper container, liquid mainly has
The dangerous liquids such as the safe liquids such as water, beverage and petroleum have preferably plastic containers based on the characteristic of reflective Terahertz
Penetrability, the container that the application uses is plastic containers.
Such as Fig. 2, the Terahertz transceiver 12 in Liquid identification device provided by the present application issues THz wave THZ, too
Hertz wave THZ touches container 110 first, and due to medium difference, part of it is transmitted by 110 surface reflection of container, a part,
Testing liquid 120 then is touched, to be reflected by testing liquid 120, wherein echo is carried about in container and container
The information of liquid includes by the reflected THZ1 of container outer wall and the THZ2 by reflection liquid back, Terahertz transceiver
12 further receive the echo being reflected back, to get the first time-domain signal map as shown in Figure 3.
S12 carries out pretreatment to the first time-domain signal map and obtains the second time-domain signal map.
After getting the first time-domain signal map, denoising and interception further are carried out to the first time-domain signal map
Peak value processing, to obtain the higher second time-domain signal map of signal-to-noise ratio.
Referring to Fig. 4, Fig. 4 is the second embodiment flow diagram of the method for the application Liquid identification, it is Fig. 1 step
The sub-step of S12, specifically comprises the following steps:
S121 removes the Gaussian noise in the first time-domain signal map by orthogonal wavelet transformation method；When obtaining third
Domain signal map.
Firstly, the Gaussian noise in the first time-domain signal map is removed by orthogonal wavelet transformation method, to obtain opposite
The higher third time-domain signal map of signal-to-noise ratio,
S122 carries out interceptionization processing to third time-domain signal map, to obtain the second time-domain signal map.
Further, the maximum maximum reflection peak information of peak value in the third time-domain signal map is obtained, is specifically existed
The location information at the maximum peak of peak value in third time-domain signal map, then according to maximum reflection peak information in third time-domain signal
The secondary maximum peak information of peak value second is obtained in map；I.e. by maximum reflection peak exclude after remaining third time domain believe
The maximum reflection peak information of peak value is obtained in number map after the acquisition to intercept secondary maximum peak information, i.e., it is this time is maximum
Peak is extracted from third time-domain signal map, to obtain the second time-domain signal map.
Due in collection process, the first time-domain signal map obtained include container reflection signal and liquid it is anti-
Signal is penetrated, and it is desirable that the reflection signal of liquid, i.e. its secondary reflection signal, first time (being reflected by container) are peak
It is worth maximum peak, second is time maximum peak, therefore by the above method, has got the time-domain signal figure of carrying of liquids information
Spectrum.
In a particular embodiment, for certain higher liquid of reflectivity, the maximum peak of peak value may be liquid
The peak of reflection, therefore available peak value ranks the first the peak with second, and by the peak in time relationship rearward intercept from
And obtain the second time-domain signal map.
S13 carries out characteristic processing to the second time-domain signal map to obtain the main feature vector of the second time-domain signal map.
After getting the higher second time-domain signal map of signal-to-noise ratio, further the second time-domain signal map is carried out special
Signization processing, to obtain the main feature vector of the second time-domain signal map.
Referring to Fig. 5, Fig. 5 is the 3rd embodiment flow diagram of the application Liquid identification method, it is Fig. 1 step
The sub-step of S13, specifically comprises the following steps:
Second time-domain signal map is carried out Fast Fourier Transform (FFT) to obtain the first frequency-region signal frequency spectrum by S131.
After second time-domain signal map is carried out Fast Fourier Transform (FFT), the first frequency-region signal frequency spectrum, relatively second are obtained
For time-domain signal map, the Mathematics structural of the first frequency-region signal frequency spectrum is become apparent, can be intuitively anti-by mathematic(al) structure
Answer the characteristic information of its second time-domain signal map.
The first frequency-region signal frequency spectrum is normalized to obtain the second frequency-region signal frequency spectrum in S132.
Then the first frequency-region signal frequency spectrum is normalized to obtain the second frequency-region signal frequency spectrum, finds out first
The variance of one frequency-region signal frequency spectrum and the numerical value of mean value.
It can then be handled according to following formula:
Wherein, yiFor the first frequency-region signal frequency spectrum, μ is the mean value of the first frequency-region signal frequency spectrum, and the δ is the first frequency domain letter
The variance of number frequency spectrum, the x (n) are the second frequency-region signal frequency spectrum, the numerical value of the second obtained frequency-region signal frequency spectrum [- 1,
1] between.
In a particular embodiment, before being acquired, metallic mirror can also be carried out using Terahertz transceiver
It acquires and handles, obtain a benchmark signal.To the benchmark signal as normalized.
S133, to the second frequency-region signal frequency spectrum carry out principal component analysis with extract the second main feature of frequency-region signal frequency spectrum to
Amount.
Principal component analysis (Principal Component Analysis, PCA) is a kind of common Data Dimensionality Reduction, feature
Selection method, it can eliminate existing correlation between data, and then filter out from multidimensional data big to classification contribution rate
Feature, reduce calculation amount, promoted discrimination.
The covariance matrix C and mean value x of the second frequency-region signal frequency spectrum are obtained first；Then obtain the spy of covariance matrix C
Value indicative D and feature vector V；Arranged according to the size of characteristic value D feature vector V corresponding to characteristic value D, specifically according to
The size of characteristic value D is arranged by row from top to bottom, and the corresponding k dimensional feature vector V composition of k characteristic value D is special before then taking
Levy matrix P.
Wherein, as follows about the confirmation formula of the value of k:
Wherein, the DiThe ith feature value of the matrix after descending arrangement, and k≤n are carried out for covariance matrix C；And k
For positive integer.
Further, by being projected the second frequency-region signal frequency spectrum x (n) to eigenmatrix P to obtain main feature vector
F.Shown in the following formula of projection process:
Wherein, F is main feature vector, and x (n) is the second frequency-region signal frequency spectrum, and x is the above-mentioned mean value found out, and P is characterized square
Battle array.
After above-mentioned principal component analysis, can according to the size of contribution degree by the second frequency-region signal frequency spectrum x (n) more
Important characteristic information extracts, and after being projected, to obtain main feature vector.
Main feature vector is inputted default classifier and classified by S14.
After getting main feature vector, further main feature vector is input in default classifier and is classified, this
The default classifier that application provides is that there is the feature vector of multiple groups liquid to be trained in support vector machine classifier (SVM)
And the sorter model obtained, in embodiment, SVM classifier is using radial kernel function.
SVM (Support Vector Machine) classifier is a kind of machine learning class classifier based on SVM method,
The principle of its SVM method be it be linear can a point situation analyzed, the case where for linearly inseparable, by using non-
The sample of low-dimensional input space linearly inseparable is converted high-dimensional feature space by Linear Mapping algorithm makes its linear separability, thus
Make it possible that high-dimensional feature space carries out linear analysis using nonlinear characteristic of the linear algorithm to sample.And it is based on knot
Optimal hyperlane is constructed in feature space on structure risk minimization theory, so that learner obtains global optimization, and
The expectation of entire sample space meets certain upper bound with some probability.I.e. it has better adaptivity and Generalization Capability, energy
It is enough preferably to handle nonlinear problem.
It specifically can be by plastic containers and different liquid, such as liquid common on the market, such as water, beverage peace
It is respectively combined to be formed to be used as initial training sample by full liquid, the dangerous liquids such as gasoline, DDVP and its mixing liquid
This, specifically extracts the main feature vector of these combination liquid as training group, it may include multistage label, such as level-one mark
Label can be with safety, including safety and danger, and second level label can be type, such as the beverage type liquid or edible oil under safely
Deng can further include the milk in three-level label, such as beverage type liquid, by the corresponding liquid of its training group and its each
Label corresponding to liquid is input to multiple SVM classifiers and is trained, and then determines that it is more by the method for cross validation
The identification accuracy of a SVM classifier model, and using it is optimal that as default classifier.
In other embodiments, deep-neural-network, deep-neural-network (Deep Neural can also be used
Network, DNN) it is the neural network with multiple hidden layers, compared to traditional classifier, it has large-scale parallel
Distributed frame, thus model has better adaptivity and Generalization Capability, can preferably handle nonlinear problem, equally
The main feature vector of aforesaid liquid group and its corresponding label are input to deep-neural-network to be trained to obtain classifier mould
Type, and confirm by cross validation the identification accuracy etc. of multiple sorter models, and using it is optimal that as default point
Class device.
S15 obtains default classifier classification results, to obtain the type of liquid to be detected.
Therefore it after getting main feature vector, is inputted default classifier and classifies, then the available main spy
The recognition result of liquid to be detected corresponding to vector is levied, recognition result includes safety and type, as its safety includes
Safety is dangerous, and type includes its specific classification, such as beverage class, or can be further specific to herbal tea.
Further, also it is known that the type of liquid to be detected, if be safe class I liquid I, or dangerous class I liquid I.Such as
Fruit its under safe class I liquid I, may further obtain it is the herbal tea in beverage class.
In a particular embodiment, default classifier provided by the present application can also be trained again at work, thus
Reinforce its recognition capability and accuracy.
Referring to Fig. 6, Fig. 6 is the first embodiment flow diagram of the application feature extracting method, specifically include as
Lower step:
S21 acquires the first time-domain signal map of liquid.
The the first time-domain signal map for acquiring liquid (stores in a reservoir) the first terahertz signal of transmitting to liquid,
And receive the second terahertz signal returned by reflection liquid.
S22 carries out pretreatment to the first time-domain signal map and obtains the second time-domain signal map.
After getting the first time-domain signal map, denoising and interception further are carried out to the first time-domain signal map
Peak value processing, to obtain the higher second time-domain signal map of signal-to-noise ratio.
S23 carries out characteristic processing to the second time-domain signal map to obtain the feature vector of the second time-domain signal map.
After getting the higher second time-domain signal map of signal-to-noise ratio, further the second time-domain signal map is carried out special
Signization processing, to obtain the feature vector of the second time-domain signal map.
The specific extracting method of the present embodiment has had narration in the above-described embodiments, and which is not described herein again, and needs to know
Road, the extracted feature vector of the feature extracting method of the present embodiment are the main feature vector in above-described embodiment,
It can be applied to the recognition methods in any of the above-described embodiment.
Referring to Fig. 7, Fig. 7 is an embodiment structure schematic block diagram of the application Liquid identification device.
Liquid identification device provided in this embodiment specifically includes processor 10, memory 11, Terahertz transceiver 12,
In, processor 10 connects memory 11 and Terahertz transceiver 12.
In the present embodiment, Terahertz transceiver 12 is used to acquire the first time-domain signal map of testing liquid.
Wherein, processor 10 can also be known as CPU (Central Processing Unit, central processing unit).Processing
Device 10 may be a kind of IC chip, the processing capacity with signal.Processor 10 can also be general processor, number
Signal processor (DSP), specific integrated circuit (ASIC), ready-made programmable gate array (FPGA) or other programmable logic devices
Part, discrete gate or transistor logic, discrete hardware components.General processor can be microprocessor or the processor
It is also possible to any conventional processor etc..
In this embodiment, processor 10 can be used for carrying out the first time-domain signal map pretreatment acquisition the second time domain letter
Number map；Characteristic processing is carried out to obtain the main feature vector of the second time-domain signal map to the second time-domain signal map；It will lead
Feature vector inputs default classifier and classifies；The default classifier classification results are obtained, to obtain the liquid to be detected
The type of body.
The modular terminal of above equipment can specifically execute step corresponding in above method embodiment respectively, therefore not right herein
Each module is repeated, and please refers to the explanation of the above corresponding step in detail.
It is the structural schematic diagram of one embodiment of the application storage device refering to Fig. 8, Fig. 8, can be realized above-mentioned all
The command file 21 of method, the command file 21 can be stored in the form of software products in above-mentioned storage device, simultaneously also
It is to record the data of various calculating, including some instructions are used so that a computer equipment (can be personal computer, service
Device, intelligent robot or network equipment etc.) or processor (processor) the execution each embodiment method of the application
All or part of the steps.
Described instruction file 21 also have certain independence, can when operating system, standby system break down after
Continuous cooperation processor 10 executes dependent instruction, will not be replaced, damage and clearly in upgrading, bootstrap upgrading and in repairing
It is empty.
And storage device above-mentioned include: USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), with
Machine accesses various Jie that can store program code such as memory (RAM, Random Access Memory), magnetic or disk
The terminal devices such as matter or computer, server, mobile phone, plate.
In conclusion first time-domain signal map of the application by the acquisition testing liquid of acquisition testing liquid, then
Pretreatment is carried out with characteristic processing to obtain its main feature vector to the first time-domain signal map, and main feature vector is inputted
Classify to default classifier, to obtain the type of testing liquid.Pass through the reflection to Terahertz using different liquids
Performance is different, and the first obtained time-domain signal map is also different, to extract the letter of the first time domain corresponding to testing liquid
The main feature vector of number map, and default classifier is provided to carry out identification calculating to main feature vector, to obtain prepare liquid
The type of body, can reduced cost significantly, and recognition speed is high, and recognition accuracy is high.
The above is only presently filed embodiments, are not intended to limit the scope of the patents of the application, all to utilize the application
Equivalent result made by specification and accompanying drawing content or equivalent process transformation are applied directly or indirectly in other relevant technologies
Field similarly includes in the scope of patent protection of the application.
Claims (11)
1. a kind of Liquid identification method based on reflective Terahertz, which is characterized in that the described method includes:
The first time-domain signal map of testing liquid in a reservoir is placed in acquisition；
Pretreatment is carried out to the first time-domain signal map and obtains the second time-domain signal map；
Characteristic processing is carried out to obtain the main feature vector of the second time-domain signal map to the second time-domain signal map；
The main feature vector is inputted default classifier to classify；
The default classifier classification results are obtained, to obtain the type of the liquid to be detected.
2. recognition methods according to claim 1, which is characterized in that described to be carried out in advance to the first time-domain signal map
Processing obtains the second time-domain signal map
The Gaussian noise in the first time-domain signal map is removed by orthogonal wavelet transformation method；To obtain third time-domain signal
Map；
Interceptionization processing is carried out to the third time-domain signal map, to obtain the second time-domain signal map.
3. recognition methods according to claim 2, which is characterized in that described to be cut to the third time-domain signal map
Taking is handled, and includes: to obtain the second time-domain signal map
Obtain the maximum maximum reflection peak information of peak value in the third time-domain signal map；
The secondary maximum peak information of peak value second is obtained in the third time-domain signal map according to maximum reflection peak information；
The secondary maximum peak information is extracted, to obtain the second time-domain signal map.
4. recognition methods according to claim 1, which is characterized in that described to carry out spy to the second time-domain signal map
Sign is handled with the main feature vector for obtaining the second time-domain signal map
The second time-domain signal map is subjected to Fast Fourier Transform (FFT) to obtain the first frequency-region signal frequency spectrum；
The first frequency-region signal frequency spectrum is normalized to obtain the second frequency-region signal frequency spectrum；
Principal component analysis is carried out to extract the main feature vector of the second frequency-region signal frequency spectrum to the second frequency-region signal frequency spectrum.
5. recognition methods according to claim 4, which is characterized in that described to return to the first frequency-region signal frequency spectrum
One, which changes processing to obtain the second frequency-region signal frequency spectrum, includes:
It is normalized according to the following formula to the first frequency-region signal frequency spectrum to obtain the second frequency-region signal frequency spectrum:
Wherein, the yiFor the first frequency-region signal frequency spectrum, the μ is the mean value of the first frequency-region signal frequency spectrum, and the δ is the first frequency
The variance of domain signal spectrum, the x (n) are the second frequency-region signal frequency spectrum.
6. recognition methods according to claim 4, which is characterized in that described to be led to the second frequency-region signal frequency spectrum
Constituent analysis includes: to extract the main feature vector of the second frequency-region signal frequency spectrum
Obtain the covariance matrix and mean value of the second frequency-region signal frequency spectrum；
Obtain the eigen vector of the covariance matrix；
The corresponding described eigenvector of the characteristic value is arranged according to the size of the characteristic value, and takes its preceding k a most
K dimensional feature vector corresponding to big characteristic value is formed to obtain eigenmatrix；
The value of the k is found out according to the following formula；
Wherein, the DiThe ith feature value of matrix after being arranged for descending, and k≤n；And k is positive integer；
The second frequency-region signal frequency spectrum is projected according to the following formula to obtain the main feature vector to eigenmatrix:
Wherein, the F is the main feature vector, and the x (n) is the second frequency-region signal frequency spectrum, describedIt is described equal
Value, the P are the eigenmatrix.
7. recognition methods according to claim 1, which is characterized in that testing liquid in a reservoir is placed in the acquisition
First time-domain signal map includes:
THz wave is emitted to the testing liquid of the placement in a reservoir；
Obtain the echo reflected via the container and the testing liquid；
Using the map of the echo as the first time-domain signal map of the testing liquid.
8. Liquid identification method according to claim 1, which is characterized in that the default classifier is by will acquire
Preset quantity multiple groups liquid main feature vector be trained in support vector machine classifier or deep-neural-network and
The sorter model of acquisition.
9. a kind of method that the fluid characteristics based on reflective Terahertz are extracted, which is characterized in that the described method includes:
Acquire the first time-domain signal map of liquid；
Pretreatment is carried out to the first time-domain signal map and obtains the second time-domain signal map；
Characteristic processing is carried out to obtain the feature vector of the second time-domain signal map to the second time-domain signal map；
Wherein, described eigenvector is the described in any item main feature vectors of claim 1-8.
10. a kind of Liquid identification device, which is characterized in that the Liquid identification device includes processor, memory and terahertz
Hereby transceiver, the processor connect the memory and the Terahertz transceiver；
Wherein, the Terahertz transceiver is used to acquire the first time-domain signal map for placing testing liquid in a reservoir；
The processor is used to carry out the first time-domain signal map pretreatment to obtain the second time-domain signal map；To described
Second time-domain signal map carries out characteristic processing to obtain the main feature vector of the second time-domain signal map；By the main spy
Sign vector inputs default classifier and classifies；The default classifier classification results are obtained, to obtain the liquid to be detected
Type.
11. a kind of storage device, which is characterized in that the storage device includes being able to achieve any one of claim 1-10 right
It is required that program file.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811150314.4A CN109406445A (en) | 2018-09-29 | 2018-09-29 | Liquid identification method, feature extracting method, Liquid identification device and storage device |
| PCT/CN2018/115080 WO2020062471A1 (en) | 2018-09-29 | 2018-11-12 | Liquid identifying method, feature extraction method, liquid identifying device and storage device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811150314.4A CN109406445A (en) | 2018-09-29 | 2018-09-29 | Liquid identification method, feature extracting method, Liquid identification device and storage device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109406445A true CN109406445A (en) | 2019-03-01 |
Family
ID=65465645
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811150314.4A Pending CN109406445A (en) | 2018-09-29 | 2018-09-29 | Liquid identification method, feature extracting method, Liquid identification device and storage device |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN109406445A (en) |
| WO (1) | WO2020062471A1 (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110044839A (en) * | 2019-04-25 | 2019-07-23 | 华太极光光电技术有限公司 | The detection method and detection system of bottling liquid |
| CN112485218A (en) * | 2020-11-05 | 2021-03-12 | 电子科技大学中山学院 | Terahertz dangerous liquid identification method based on artificial neural network |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114088656A (en) * | 2020-07-31 | 2022-02-25 | 中国科学院上海高等研究院 | Terahertz spectrum substance identification method and system, storage medium and terminal |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170336259A1 (en) * | 2016-05-20 | 2017-11-23 | Hamamatsu Photonics K.K. | Total reflection spectroscopic measurement device and total reflection spectroscopic measurement method |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9261456B2 (en) * | 2011-01-25 | 2016-02-16 | University Of Washington Through Its Center For Commercialization | Terahertz spectroscopy of rough surface targets |
| CN103941254A (en) * | 2014-03-03 | 2014-07-23 | 中国神华能源股份有限公司 | Soil physical property classification recognition method and device based on geological radar |
| CN104897605B (en) * | 2015-06-16 | 2018-01-23 | 中国人民解放军国防科学技术大学 | It is a kind of that classifying identification method is composed based on the Terahertz for improving SVMs |
| CN106645014B (en) * | 2016-09-23 | 2019-04-30 | 上海理工大学 | Substance identification based on tera-hertz spectra |
-
2018
- 2018-09-29 CN CN201811150314.4A patent/CN109406445A/en active Pending
- 2018-11-12 WO PCT/CN2018/115080 patent/WO2020062471A1/en active Application Filing
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170336259A1 (en) * | 2016-05-20 | 2017-11-23 | Hamamatsu Photonics K.K. | Total reflection spectroscopic measurement device and total reflection spectroscopic measurement method |
Non-Patent Citations (4)
| Title |
|---|
| JUNXIU LIU ET AL.: "Organic Compound Identification Based on Terahertz Spectrum", 《2018 IEEE 7TH DATA DRIVEN CONTROL AND LEARNING SYSTEMS CONFERENCE》 * |
| 刘畅等: "利用太赫兹反射式时域光谱测量有机溶剂光学参数", 《光谱学与光谱分析》 * |
| 姜代红著: "《复杂环境下监控图像拼接与识别》", 28 February 2017, 中国矿业大学出版社 * |
| 张文涛等: "基于太赫兹时域光谱技术与PCA-SVM 的转基因大豆油鉴别研究", 《红外与激光工程》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110044839A (en) * | 2019-04-25 | 2019-07-23 | 华太极光光电技术有限公司 | The detection method and detection system of bottling liquid |
| CN112485218A (en) * | 2020-11-05 | 2021-03-12 | 电子科技大学中山学院 | Terahertz dangerous liquid identification method based on artificial neural network |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2020062471A1 (en) | 2020-04-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109406445A (en) | Liquid identification method, feature extracting method, Liquid identification device and storage device | |
| US11631152B2 (en) | Security check system and method for configuring security check device | |
| CN109470720B (en) | Liquid identification method, vector extraction method, liquid identification device, and storage medium | |
| CN105260437B (en) | Text classification feature selection approach and its application in biological medicine text classification | |
| CN112990942B (en) | Cloud computing-based intelligent chemical storage management system and method | |
| Ergen et al. | Texture based feature extraction methods for content based medical image retrieval systems | |
| CN101667246A (en) | Human face recognition method based on nuclear sparse expression | |
| de Lima et al. | Methods of authentication of food grown in organic and conventional systems using chemometrics and data mining algorithms: A review | |
| CN107358344A (en) | Enterprise's hidden danger management method and its management system, electronic equipment and storage medium | |
| CN109684476A (en) | A kind of file classification method, document sorting apparatus and terminal device | |
| Sakaguchi et al. | Recognizing subsurface target responses in ground penetrating radar data using convolutional neural networks | |
| CN109374572A (en) | Terahertz time-domain spectroscopy taxonomy of goods method neural network based | |
| Reichman et al. | Improvements to the Histogram of Oriented Gradient (HOG) prescreener for buried threat detection in ground penetrating radar data | |
| CN113642646B (en) | Image threat object classification and positioning method based on multi-attention and semantics | |
| Rather et al. | Evaluation of machine learning models for a chipless RFID sensor tag | |
| Yang et al. | Identification of marine oil spill pollution using hyperspectral combined with thermal infrared remote sensing | |
| Palo et al. | Fault detection in seismic data using graph attention network | |
| CN109766899A (en) | Physical features extract and the SAR image vehicle target recognition methods of SVM | |
| Ferdosi et al. | Identifying counterfeit medicine in Bangladesh using deep learning | |
| CN106056339B (en) | A kind of article identification method using electromagnetic wave penetrability | |
| Parisotto et al. | Unsupervised clustering of Roman pottery profiles from their SSAE representation | |
| CN109409417A (en) | Liquid identification method, apparatus and storage device based on heat transfer | |
| CN110045371A (en) | Identification method, device, equipment and storage medium | |
| Stump et al. | An exploration of gradient-based features for buried threat detection using a handheld ground penetrating radar | |
| Kim et al. | Optimal wavelet packets for characterizing surface quality |
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: 20190301 |
|
| RJ01 | Rejection of invention patent application after publication |