CN109241937A - Recognition methods, device and storage medium based on THz wave - Google Patents
Recognition methods, device and storage medium based on THz wave Download PDFInfo
- Publication number
- CN109241937A CN109241937A CN201811126077.8A CN201811126077A CN109241937A CN 109241937 A CN109241937 A CN 109241937A CN 201811126077 A CN201811126077 A CN 201811126077A CN 109241937 A CN109241937 A CN 109241937A
- Authority
- CN
- China
- Prior art keywords
- neural network
- network model
- data
- output
- thz wave
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 28
- 238000003062 neural network model Methods 0.000 claims abstract description 46
- 239000011159 matrix material Substances 0.000 claims description 35
- 238000013528 artificial neural network Methods 0.000 claims description 9
- 230000009467 reduction Effects 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 5
- 230000004888 barrier function Effects 0.000 claims description 4
- 238000000513 principal component analysis Methods 0.000 claims description 4
- 241001269238 Data Species 0.000 claims description 2
- 210000004218 nerve net Anatomy 0.000 claims 1
- 238000013527 convolutional neural network Methods 0.000 description 11
- 238000010276 construction Methods 0.000 description 3
- 238000003384 imaging method Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000001537 neural effect Effects 0.000 description 2
- 230000005855 radiation Effects 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
- G06F2218/10—Feature extraction by analysing the shape of a waveform, e.g. extracting parameters relating to peaks
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Immunology (AREA)
- General Health & Medical Sciences (AREA)
- Biochemistry (AREA)
- Analytical Chemistry (AREA)
- Pathology (AREA)
- Chemical & Material Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Signal Processing (AREA)
- General Engineering & Computer Science (AREA)
- Toxicology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
- Image Analysis (AREA)
Abstract
The present invention discloses a kind of recognition methods based on THz wave, device and storage medium.Wherein, method includes obtaining different determinands；The determinand is placed respectively under THz wave background and obtains the corresponding terahertz time-domain Wave data of the determinand；The amplitude points of different frequency component are obtained according to the terahertz time-domain Wave data；Initial data is constructed according to the amplitude points；Neural network model input of several initial data as a configuration is selected, for obtaining the output of the neural network model；According to the output of the neural network model, the determinand is identified.The present invention can identify the object under THz wave background by constructing neural network model.
Description
Technical field
The present invention relates to rote learning field, in particular to a kind of recognition methods based on THz wave, device and
Storage medium.
Background technique
Multifrequency Terahertz detection system is referred in the open source literature of CN104965233A.
Since the THz wave of different frequency is after the transmission of testee and/or scattering and/or absorbing, Strength Changes
It is different；The detector that so system passes through different frequency obtains the terahertz light of Strength Changes at different frequencies respectively
After generate electric signal, then be imaged by the electric signal.The imaging of system is implemented for the identification to testee；
It is usually artificial judgment to the identification of testee in such as above-mentioned imaging or passes through figure under the imaging of high contrast
As technology is realized.
Summary of the invention
The embodiment of the present invention at least provides a kind of recognition methods based on THz wave, can be by constructing neural network mould
Type identifies the object under THz wave background.
The specific implementation of above-described embodiment, as described below.
The described method includes:
Obtain different determinands；
The determinand is placed respectively under THz wave background and obtains the corresponding Terahertz frequency domain of the determinand
Wave data；
The amplitude points of different frequency component are obtained according to the terahertz time-domain Wave data；
Initial data is constructed according to the amplitude points；
Neural network model input of several initial data as a configuration is selected, for obtaining the neural network
The output of model；
According to the output of the neural network model, the determinand is identified.
In some embodiments disclosed by the invention, obtains the terahertz time-domain Wave data and be configured to；
The terahertz time-domain Wave data of the determinand is obtained under the THz wave background of a frequency point；
Fast Fourier Transform (FFT) is carried out to the terahertz time-domain Wave data, obtains the Terahertz frequency wave figurate number
According to.
In some embodiments disclosed by the invention, several initial data of the selection are as a neural network mould
Type input, is configured that
Primitive character matrix is constructed according to several initial data；
Principal Component Analysis Algorithm is selected to obtain the dimensionality reduction matrix of the primitive character matrix；
The drop matrix is selected to input as the neural network model.
In some embodiments disclosed by the invention, configuring the neural network model includes:
It obtains different samples and marks the label of the different samples；
The terahertz time-domain Wave data of the sample is obtained respectively；
The sample point of different frequency component is obtained according to the terahertz time-domain Wave data；
Sample data is constructed according to the sample point；
Sample characteristics matrix is constructed according to several sample datas；
Principal Component Analysis Algorithm is selected to obtain the sample dimensionality reduction matrix of the sample characteristics matrix；
The sample drop matrix is selected to input as the neural network model and select the label as the mind
It is exported through network model；
According to the sample, matrix and the label training neural network model drop.
In some embodiments disclosed by the invention, the training neural network model is configured that
Configure the input layer, hidden layer and output layer of the neural network model；
The weight of random arrangement input layer, hidden layer and output layer；
Input of the sample drop matrix as the neural network model is selected, for obtaining the neural network model
Reality output；
To the pass-algorithm reality output and the label after selection, for updating in the neural network model
The weight of input layer, hidden layer and output layer.
In some embodiments disclosed by the invention, it is less than minimum threshold in the reality output and the deviation of label
Stop updating the weight afterwards.
In some embodiments disclosed by the invention, absorptivity, refraction are obtained according to the terahertz time-domain Wave data
Rate and transmissivity；
The initial data, the absorptivity, the refractive index and the transmissivity is combined to combine as initial data；
The neural network model input that several original data set cooperations are a configuration is selected, for obtaining the nerve
The output of network model；
According to the output of the neural network model, the determinand is identified.
The embodiment of the present invention at least discloses a kind of storage medium, for storing computer instruction, which is characterized in that the finger
It realizes when order is executed by processor such as the step of the above method.
The embodiment of the present invention at least discloses a kind of identification device based on THz wave, and described device includes:
Data acquisition module is configured to acquisition and is placed on the corresponding Terahertz frequency domain of determinand under THz wave background
Wave data obtains the amplitude points of different frequency component according to the terahertz time-domain Wave data, according to the amplitude points structure
Build initial data；
Neural network module identifies the determinand according to several initial data.
The embodiment of the present invention at least discloses a kind of recognition methods based on THz wave,
The described method includes:
Obtain the combination of determinand and any barrier；
The combination is placed respectively under THz wave background and obtains the corresponding Terahertz frequency-domain waveform of combination
Data；
The amplitude points of different frequency component are obtained according to the terahertz time-domain Wave data；
Initial data is constructed according to the amplitude points；
Neural network model input of several initial data as a configuration is selected, for obtaining the neural network
The output of model；
According to the output of the neural network model, the combination is identified.
For above scheme, the present invention is by being referring to the drawings described in detail disclosed exemplary embodiment, also
The other feature and its advantage for making the embodiment of the present invention understand.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below will be to needed in the embodiment attached
Figure is briefly described, it should be understood that the following drawings illustrates only certain embodiments of the present invention, therefore is not construed as pair
The restriction of range for those of ordinary skill in the art without creative efforts, can also be according to this
A little attached drawings obtain other relevant attached drawings.
Fig. 1 is the flow chart of embodiment one；
Fig. 2 is the flow chart of embodiment two；
Fig. 3 is the schematic diagram of two application scenarios of embodiment.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.The present invention being usually described and illustrated herein in the accompanying drawings is implemented
The component of example can be arranged and be designed with a variety of different configurations.
Therefore, the detailed description of the embodiment of the present invention provided in the accompanying drawings is not intended to limit below claimed
The scope of the present invention, but be merely representative of selected embodiment of the invention.Based on the embodiments of the present invention, this field is common
Technical staff's every other embodiment obtained without creative efforts belongs to the model that the present invention protects
It encloses.
Embodiment 1
The present embodiment provides a kind of recognition methods based on THz wave.Method through this embodiment can be according to building
Convolutional neural networks model determinand different under THz wave background is identified.
Referring to FIG. 1, the present embodiment is as follows to the recognition methods of determinand.
Step100, several different determinands are obtained.
Step210, the THz wave background that determinand is individually positioned in a frequency point test environment under；And it is testing
Corresponding terahertz time-domain Wave data is obtained under environment.
Step220, Fast Fourier Transform (FFT) is carried out to the terahertz time-domain Wave data of acquisition, obtains the Terahertz frequency
Rate Wave data.
Step310, all frequency components for obtaining terahertz time-domain Wave data.
Step320, in view of the frequency range of THz wave is in 100Ghz to 10Thz；Frequency component is in frequency spectrum coordinate system
In △ X be approximately 0.All frequency component essence so obtained are amplitude points.
The amplitude points that Step400, the present embodiment setting obtain have m, if the amplitude points obtained according to determinand are less than m
It is a, then it is supplied with 0.Characteristic point of the m amplitude points as an initial data.N initial data constructs primitive character matrix X；Its
In,
Step511, the covariance matrix C for calculating primitive character matrix X.
General covariance formula is,It is characterized mean value.
X, Y sample and covariance formula be timing, illustrate X and Y be positive correlation；When covariance formula is negative, say
Bright X and Y is negative correlativing relation, and X and Y is mutually indepedent when covariance is 0.
In view of the data characteristics of raw data matrix X, the present embodiment simplifies covariance, specific as follows.
Step512, the characteristic mean for obtaining all characteristic points.
Step513, zero averaging processing is carried out according to all characteristic points of the characteristic mean to primitive character matrix X, even ifAndIt is 0.
So simplified covariance formula is,
Step514, due to primitive character matrix X, covariance is symmetry square matrix.Then covariance matrix C is,
Step520, the present embodiment carry out singular value decomposition to covariance square C, calculate characteristic value and the spy of covariance matrix C
Levy vector.
Step530, dimensionality reduction matrix P is constructed according to characteristic value and feature vector；
Step540, sequential arrangement is carried out to the corresponding feature vector of characteristic value according to the size of characteristic value；
Step550, eigenvectors matrix Z is established according to the feature vector of arrangement；
Step560, selected characteristic vector matrix Z preceding K row construct dimensionality reduction matrix P, K be less than N and be positive integer.
Step570, the product for calculating primitive character matrix X and dimensionality reduction matrix X, i.e. PX is as the number after initial data dimensionality reduction
According to collection T, the data set T that the present embodiment obtains is the dimensionality reduction data of initial data.
Step600, it is inputted according to data set T as the convolutional neural networks that are pre-created, for identification determinand.
The building configuration of the present embodiment convolutional neural networks is as follows.
Step110, it obtains different samples and label is carried out to different samples.
The step of Step120, such as the present embodiment, obtains the data set Tx of sample.
Step130, it is based on MATLAB platform construction convolutional neural networks model；Select data set Tx as convolutional Neural net
The standard of network model inputs；Select standard output of the label as convolutional neural networks model.
The nodes such as Step140, input layer, hidden layer and the output layer for configuring convolutional neural networks model.
The weight of Step150, random configuration arbitrary node.
Step160, according to standard input and standard output to convolutional neural networks model training；And to transmitting after selecting
Algorithm according to reality output compared with the output of configuration, repeatedly update convolutional neural networks in all nodes weight.
Step170, the reality output that convolutional neural networks are obtained according to standard input is obtained；Compare reality output and standard
The deviation of output, and be less than minimum threshold that is, after the convergence of convolutional neural networks structural parameters in deviation and complete convolutional Neural net
The building of network model stops updating weight.
Preferably, it is as follows to be based on MATLAB platform construction convolutional neural networks model for the present embodiment.
Configure the number of convolutional layer and the characteristic pattern number and size of the application of all convolutional layers；Convolutional layer A shares 16 and goes
A feature vector, each feature vector are deconvoluted using one 11 × 1 convolution kernel, and moving step length is set as 2, which exports 32
The feature vector of 93 × 1 sizes；Convolutional layer B uses 9 × 1 convolution kernel, exports the feature vector of 32 85 × 1 sizes；Chi Hua
Layer a uses 3 × 1 Chi Huahe, and moving step length 2 generates the feature vector of 32 42 × 1 sizes；Convolutional layer C is using 7 × 1
Convolution kernel, moving step length 2 export the convolution kernel of 32 18 × 1 sizes；Convolutional layer D uses 5 × 1 convolution kernel, exports 32
The feature vector of 14 × 1 sizes；Pond layer b uses 3 × 1 Chi Huahe, and moving step length 2 generates the spy of 32 6 × 1 sizes
Levy vector；Convolutional layer E uses 6 × 1 convolution kernel, exports the feature vector of 32 1 × 1 sizes；
The present embodiment separately discloses a kind of storage medium, for ease of description, illustrates only relevant to the embodiment of the present invention
Part.Storage medium is for storing at least one computer instruction, realization when at least one computer instruction is executed by processor
The all or part of the steps of the above method.
The present embodiment separately discloses a kind of identification device based on THz wave and illustrates only and this hair for ease of description
The relevant part of bright embodiment.Device includes data acquisition module and neural network module.
The data acquisition module of the present embodiment, which is configured to obtain, is placed on the corresponding terahertz of determinand under THz wave background
Hereby frequency-domain waveform data obtain the amplitude points of different frequency component according to terahertz time-domain Wave data, are constructed according to amplitude points
Initial data；
The neural network module of the present embodiment is configured to identify determinand according to several initial data.
Embodiment two
Referring to FIG. 2, the present embodiment applies the method for embodiment one using human body as the scene of THz wave background, such as
The article that Fig. 3 hides under the clothing human body detects.
The method of the present embodiment includes:
Step100, the combination for obtaining determinand and different barriers；
Step200, respectively place combination simulation human body radiation THz wave environment under, i.e., with the terahertz of human body radiation
Hereby wave frequency point is close；The corresponding terahertz time-domain Wave data of combination is obtained again.
Step300, the amplitude points that different frequency component is obtained according to terahertz time-domain Wave data.
Step400, according to amplitude points as characteristic point construct initial data.
Step500, neural network model input of several initial data as a configuration is selected, it is described for obtaining
The output of neural network model；The building of neural network model can select embodiment first is that model construction is implemented.
Step600, the output according to neural network model, identify different combinations.
Through the above scheme, the present embodiment can identify the combination of different barriers and determinand, assist safety check
Personnel check whether test object carries contraband.
Embodiment three
Absorptivity, refractive index and the transmissivity that the present embodiment is obtained according to terahertz time-domain Wave data make building as special
Point is levied, simple initial data is constructed, for quickly identifying determinand.
The present embodiment concrete configuration is to obtain absorptivity, refractive index and transmissivity according to terahertz time-domain Wave data.This
Embodiment selects several absorptivities, refractive index and transmissivity combination as the neural network model input of a configuration, for obtaining
The output of neural network model.
Example IV
The present embodiment is with respect to absorptivity, refractive index and the transmissivity that embodiment one is obtained according to terahertz time-domain Wave data
Make to construct relative complex initial data combination, for more accurately identifying determinand.The present embodiment selects several initial data
The neural network model input as a configuration is combined, for obtaining the output of neural network model；According to neural network model
Output, identify determinand.
The foregoing is only a preferred embodiment of the present invention, is not intended to restrict the invention, for the skill of this field
For art personnel, the invention may be variously modified and varied.All within the spirits and principles of the present invention, made any to repair
Change, equivalent replacement, improvement etc., should all be included in the protection scope of the present invention.
Claims (10)
1. a kind of recognition methods based on THz wave, which is characterized in that
The described method includes:
Obtain different determinands；
The determinand is placed respectively under THz wave background and obtains the corresponding Terahertz frequency-domain waveform of the determinand
Data；
The amplitude points of different frequency component are obtained according to the terahertz time-domain Wave data；
Initial data is constructed according to the amplitude points；
Neural network model input of several initial data as a configuration is selected, for obtaining the neural network model
Output；
According to the output of the neural network model, the determinand is identified.
2. as described in claim 1 based on the recognition methods of THz wave, which is characterized in that
The terahertz time-domain Wave data is obtained to be configured to；
The terahertz time-domain Wave data of the determinand is obtained under the THz wave background of a frequency point；
Fast Fourier Transform (FFT) is carried out to the terahertz time-domain Wave data, obtains the Terahertz frequency Wave data.
3. as described in claim 1 based on the recognition methods of THz wave, which is characterized in that
Several initial data of the selection are inputted as a neural network model, are configured that
Primitive character matrix is constructed according to several initial data；
Principal Component Analysis Algorithm is selected to obtain the dimensionality reduction matrix of the primitive character matrix；
The drop matrix is selected to input as the neural network model.
4. as claimed in claim 3 based on the recognition methods of THz wave, which is characterized in that
Configuring the neural network model includes:
It obtains different samples and marks the label of the different samples；
The terahertz time-domain Wave data of the sample is obtained respectively；
The sample point of different frequency component is obtained according to the terahertz time-domain Wave data；
Sample data is constructed according to the sample point；
Sample characteristics matrix is constructed according to several sample datas；
Principal Component Analysis Algorithm is selected to obtain the sample dimensionality reduction matrix of the sample characteristics matrix；
The sample drop matrix is selected to input as the neural network model and select the label as the nerve net
The output of network model；
According to the sample, matrix and the label training neural network model drop.
5. as claimed in claim 4 based on the recognition methods of THz wave, which is characterized in that
The training neural network model, is configured that
Configure the input layer, hidden layer and output layer of the neural network model；
The weight of random arrangement input layer, hidden layer and output layer；
Input of the sample drop matrix as the neural network model is selected, for obtaining the reality of the neural network model
Border output；
To the pass-algorithm reality output and the label after selection, inputted for updating in the neural network model
The weight of layer, hidden layer and output layer.
6. as claimed in claim 5 based on the recognition methods of THz wave, which is characterized in that
Stop updating the weight after the reality output and the deviation of label are less than minimum threshold.
7. as described in claim 1 based on the recognition methods of THz wave, which is characterized in that
Absorptivity, refractive index and transmissivity are obtained according to the terahertz time-domain Wave data；
The initial data, the absorptivity, the refractive index and the transmissivity is combined to combine as initial data；
The neural network model input that several original data set cooperations are a configuration is selected, for obtaining the neural network
The output of model；
According to the output of the neural network model, the determinand is identified.
8. a kind of storage medium, for storing computer instruction, which is characterized in that realized such as when described instruction is executed by processor
The step of claim 1-7 any one the method.
9. a kind of identification device based on THz wave, which is characterized in that
Described device includes:
Data acquisition module is configured to acquisition and is placed on the corresponding Terahertz frequency-domain waveform of determinand under THz wave background
Data obtain the amplitude points of different frequency component according to the terahertz time-domain Wave data, are constructed according to the amplitude points former
Beginning data；
Neural network module identifies the determinand according to several initial data.
10. a kind of recognition methods based on THz wave, which is characterized in that
The described method includes:
Obtain the combination of determinand and any barrier；
The combination is placed respectively under THz wave background and obtains the corresponding terahertz time-domain Wave data of combination；
The amplitude points of different frequency component are obtained according to the terahertz time-domain Wave data；
Initial data is constructed according to the amplitude points；
Neural network model input of several initial data as a configuration is selected, for obtaining the neural network model
Output；
According to the output of the neural network model, the combination is identified.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811126077.8A CN109241937B (en) | 2018-09-26 | 2018-09-26 | Terahertz wave-based identification method and device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811126077.8A CN109241937B (en) | 2018-09-26 | 2018-09-26 | Terahertz wave-based identification method and device and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109241937A true CN109241937A (en) | 2019-01-18 |
| CN109241937B CN109241937B (en) | 2020-01-07 |
Family
ID=65056238
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811126077.8A Active CN109241937B (en) | 2018-09-26 | 2018-09-26 | Terahertz wave-based identification method and device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109241937B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113344051A (en) * | 2021-05-28 | 2021-09-03 | 青岛青源峰达太赫兹科技有限公司 | Neural network classification method based on terahertz data |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102759753A (en) * | 2011-04-29 | 2012-10-31 | 同方威视技术股份有限公司 | Method and device for detecting hidden dangerous article |
| CN107092040A (en) * | 2017-06-01 | 2017-08-25 | 上海理工大学 | Terahertz imaging rays safety detection apparatus and video procession method |
| CN107907499A (en) * | 2017-12-28 | 2018-04-13 | 深圳市太赫兹科技创新研究院 | Multifrequency Terahertz detection device, system and method |
| CN108267420A (en) * | 2017-12-22 | 2018-07-10 | 深圳市太赫兹系统设备有限公司 | Seed aqueous amount detection systems and method |
-
2018
- 2018-09-26 CN CN201811126077.8A patent/CN109241937B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102759753A (en) * | 2011-04-29 | 2012-10-31 | 同方威视技术股份有限公司 | Method and device for detecting hidden dangerous article |
| CN107092040A (en) * | 2017-06-01 | 2017-08-25 | 上海理工大学 | Terahertz imaging rays safety detection apparatus and video procession method |
| CN108267420A (en) * | 2017-12-22 | 2018-07-10 | 深圳市太赫兹系统设备有限公司 | Seed aqueous amount detection systems and method |
| CN107907499A (en) * | 2017-12-28 | 2018-04-13 | 深圳市太赫兹科技创新研究院 | Multifrequency Terahertz detection device, system and method |
Non-Patent Citations (6)
| Title |
|---|
| JOHN F FEDERICI等: "THz imaging and sensing for security applications—explosives, weapons and drugs", 《IOPSCIENCE》 * |
| YUYING JIANG等: "Early detection of germinated wheat grains using terahertz image and chemometrics", 《SCIENTIFIC REPORTS》 * |
| 丁志龙: "太赫兹时域光谱用于混合物定性和定量分析", 《中国优秀硕士学位论文全文数据库 基础科技辑》 * |
| 周胜灵: "基于模式识别的太赫兹光谱建模及应用研究", 《中国博士学位论文全文数据库 基础科技辑》 * |
| 聂君扬等: "基于太赫兹时域光谱技术与PCA-BPN网络的转基因大豆鉴别", 《光子学报》 * |
| 陈锡爱: "太赫兹时域光谱及其成像检测技术研究", <中国博士学位论文全文数据库 基础科技辑> * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113344051A (en) * | 2021-05-28 | 2021-09-03 | 青岛青源峰达太赫兹科技有限公司 | Neural network classification method based on terahertz data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109241937B (en) | 2020-01-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Mukund et al. | Transient classification in LIGO data using difference boosting neural network | |
| Razzano et al. | Image-based deep learning for classification of noise transients in gravitational wave detectors | |
| CN107194433B (en) | A kind of Radar range profile's target identification method based on depth autoencoder network | |
| Liang et al. | CNN-based image analysis for malaria diagnosis | |
| Carrazza et al. | Lund jet images from generative and cycle-consistent adversarial networks | |
| Triastcyn et al. | Generating artificial data for private deep learning | |
| CN109325395A (en) | The recognition methods of image, convolutional neural networks model training method and device | |
| CN110163180A (en) | Mental imagery eeg data classification method and system | |
| Kartasalo et al. | CytoSpectre: a tool for spectral analysis of oriented structures on cellular and subcellular levels | |
| CN108875732A (en) | Model training and example dividing method, device and system and storage medium | |
| CN110443882A (en) | Light field microscopic three-dimensional method for reconstructing and device based on deep learning algorithm | |
| Duan et al. | Multichannel pulse-coupled neural network-based hyperspectral image visualization | |
| Bhinge et al. | Extraction of time-varying spatiotemporal networks using parameter-tuned constrained IVA | |
| Chatterjee et al. | Extraction of binary black hole gravitational wave signals from detector data using deep learning | |
| CN109325490A (en) | Terahertz image target identification method based on deep learning and RPCA | |
| Raumonen et al. | Segmentation of vessel structures from photoacoustic images with reliability assessment | |
| Ayat et al. | A comparison of artificial neural networks learning algorithms in predicting tendency for suicide | |
| Zhang et al. | On biometric verification of a user by means of eye movement data mining | |
| Nie et al. | Pulse coupled neural network based MRI image enhancement using classical visual receptive field for smarter mobile healthcare | |
| Matuszewski et al. | Recognition of electromagnetic sources with the use of deep neural networks | |
| CN105628868B (en) | A kind of undersampled signal im-pact location processing method and system of composite structure | |
| Romaniuk | Photonics applications in astronomy, communications, industry, and high-energy physics experiments 2014 | |
| Dong et al. | Multi-reservoirs EEG signal feature sensing and recognition method based on generative adversarial networks | |
| CN104573745B (en) | Fruit-fly classified method based on magnetic resonance imaging | |
| CN109241937A (en) | Recognition methods, device and storage medium based on THz wave |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |