CN110187321A - Radar emitter characteristic parameter extraction method under complex environment based on deep learning - Google Patents
Radar emitter characteristic parameter extraction method under complex environment based on deep learning Download PDFInfo
- Publication number
- CN110187321A CN110187321A CN201910462165.3A CN201910462165A CN110187321A CN 110187321 A CN110187321 A CN 110187321A CN 201910462165 A CN201910462165 A CN 201910462165A CN 110187321 A CN110187321 A CN 110187321A
- Authority
- CN
- China
- Prior art keywords
- initial characteristics
- complex environment
- characteristic parameter
- parameter
- middle layer
- 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
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
- G01S7/417—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section involving the use of neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2136—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on sparsity criteria, e.g. with an overcomplete basis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Evolutionary Biology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computer Networks & Wireless Communication (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention discloses a kind of Radar emitter characteristic parameter extraction methods under complex environment based on deep learning, belong to electronic reconnaissance field.It include: that initial characteristics extract, Classification Neural building, sparse self-encoding encoder constructs and eigenmatrix splicing；Classification Neural identification is wanted the method combined with sparse self-encoding encoder neural network recognization by the present invention, analyse in depth and study the essence of emitter Signals, explore new characteristic parameter, building is more conducive to the feature vector of signal identification, improves the recognition capability of radar emitter signal under complex environment.
Description
Technical field
The invention belongs to electronic reconnaissance fields, and in particular to Radar emitter under a kind of complex environment based on deep learning
Characteristic parameter extraction method.
Background technique
Target identification is the key link in electronic reconnaissance field, and wherein main task is that the characteristic parameter of emitter Signals mentions
It takes.Theoretical result in this respect has: instantaneous correlation method, Wavelet Transform, ambiguity function ridge characteristic method, wavelet packet and entropy feature
Method etc., their levels from different angles, different study radar target recognition.
At present Radar emitter characteristic parameter extraction there are the problem of it is as follows:
Existing method is mainly for specific signal, and in the compound complex environment of multi-signal, current
Algorithm not enough meets actual demand.
For current existing characteristic parameter and recognition methods, in the simple situation of electromagnetic environment relatively effectively, but
(such as: SNR≤2dB) recognition effect is bad under Low SNR, is not met by a variety of in-pulse modulation signals and exists simultaneously feelings
Identification requirement under condition.
The purpose of identification is to know the weapon type for emitting the signal and judge its threat level, however for radiation source
Identification, is presently considered less.
From the above analysis, further investigate the essence of emitter Signals, explore be more suitable for the feature of signal identification to
Amount is of great significance to the identification for realizing radar emitter signal under complex environment.
Summary of the invention
It is an object of the invention to: provide Radar emitter characteristic parameter under a kind of complex environment based on deep learning
Extracting method, solves under complex environment that Radar emitter characteristic parameter extraction is ineffective, and be unable to satisfy that identification requires asks
Topic.
The technical solution adopted by the invention is as follows:
A kind of Radar emitter characteristic parameter extraction method under the complex environment based on deep learning, comprising:
Initial characteristics extract: extracting the parameter information of radiation source and loading platform as initial characteristics；
Classification Neural building: input initial characteristics, building " initial characteristics-neural network middle layer A- radiation source and
Loading platform classification " upper layer Classification Neural structure exports mapping initial characteristics and radiation by neural network middle layer A
Source, loading platform class relations eigenmatrix A；
Sparse self-encoding encoder network struction: initial characteristics are used as input and output amount simultaneously, construct " initial characteristics-encoder-
The sparse self-encoding encoder network structure of neural network middle layer B- decoder " lower layer is exported initial special by neural network middle layer B
Levy the inherent attribute eigenmatrix B after being refined by depth；
Eigenmatrix splicing: it will reflect initial characteristics, radiation source and loading platform classification relationship characteristic matrix A and reflection
The eigenmatrix B of initial characteristics itself inherent attribute is stitched together, and obtains final complex environment characteristic parameter.
Further, the initial characteristics include the carrier frequency of radar, pulsewidth, angle of arrival, pulse for the parameter of radiation source
Repetition rate, antenna scan period, in conjunction with being adjusted in the pulse arrival time of signals such as communication and interference, pulse envelope parameter, arteries and veins
Parameter processed, amplitude, frequency spectrum parameter；
The initial characteristics include loading platform movement speed, space position parameter for the parameter of loading platform.
Further, the information in the Classification Neural is propagated in one direction, the neural network middle layer A's
Training method is by the way of supervised learning.
Further, in the sparse self-encoding encoder network, encoder is used to carry out dimension-reduction treatment to initial characteristics, refines
The kernel information of initial characteristics；For decoder for training encoder, whether the information for judging that encoder refines is accurate, if obtains
With the feature of initial characteristics identical information amount, and the error of output is fed back into initial characteristics, is trained among neural network with this
Layer B exports the eigenmatrix B after initial characteristics are refined by depth.
In conclusion by adopting the above-described technical solution, the beneficial effects of the present invention are:
1. the present invention is using Classification Neural and the sparse method combined from coding neural network, Classification Neural
Obtain the eigenmatrix A of mapping initial characteristics and radiation source, loading platform class relations according to initial characteristics input quantity, it is sparse from
Encoding nerve network obtains the inherent attribute eigenmatrix B after initial characteristics are refined by depth according to initial characteristics input quantity, will
Two eigenmatrixes splice to obtain final characteristic parameter, analyse in depth and study the essence of emitter Signals, explore new spy
Parameter is levied, so that building is more conducive to the feature vector of signal identification, promotes the identification of radar emitter signal under complex environment
Ability.
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, in which:
Fig. 1 is feature of present invention parameter extraction flow chart；
Fig. 2 is the sparse self-encoding encoder network diagram of the present invention；
Fig. 3 is relative entropy value variation schematic diagram.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that described herein, specific examples are only used to explain the present invention, not
For limiting the present invention, i.e., described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is logical
The component for the embodiment of the present invention being often described and illustrated herein in the accompanying drawings 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 embodiment of the present invention, those skilled in the art
Member's every other embodiment obtained without making creative work, shall fall within the protection scope of the present invention.
It should be noted that the relational terms of term " first " and " second " or the like be used merely to an entity or
Operation is distinguished with another entity or operation, and without necessarily requiring or implying between these entities or operation, there are any
This actual relationship or sequence.Moreover, the terms "include", "comprise" or its any other variant be intended to it is non-exclusive
Property include so that include a series of elements process, method, article or equipment not only include those elements, but also
Further include other elements that are not explicitly listed, or further include for this process, method, article or equipment it is intrinsic
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including described
There is also other identical elements in the process, method, article or equipment of element.
A kind of Radar emitter characteristic parameter extraction method under the complex environment based on deep learning, comprising:
Initial characteristics extract: extracting the parameter information of radiation source and loading platform as initial characteristics；
Classification Neural building: input initial characteristics, building " initial characteristics-neural network middle layer A- radiation source and
Loading platform classification " upper layer Classification Neural structure exports mapping initial characteristics and radiation by neural network middle layer A
Source, loading platform class relations eigenmatrix A；
Sparse self-encoding encoder network struction: initial characteristics are used as input and output amount simultaneously, construct " initial characteristics-encoder-
The sparse self-encoding encoder network structure of neural network middle layer B- decoder " lower layer is exported initial special by neural network middle layer B
Levy the inherent attribute eigenmatrix B after being refined by depth；
Eigenmatrix splicing: it will reflect initial characteristics, radiation source and loading platform classification relationship characteristic matrix A and reflection
The eigenmatrix B of initial characteristics itself inherent attribute is stitched together, and obtains final complex environment characteristic parameter.
Specifically, characteristic parameter extraction process is as shown in Figure 1:
Step 1: extracting initial characteristics information first: being directed to radiation source, consider carrier frequency, pulsewidth, angle of arrival, the arteries and veins of radar
The parameters such as repetition rate, antenna scan period are rushed, pulse arrival time, pulse envelope ginseng in conjunction with the signals such as communication and interference
The measurement parameters such as number, intra-pulse modulation parameter, amplitude, frequency spectrum, form initial characteristic parameter；For loading platform, then directly select
Take the features such as its movement speed, spatial position as initial characteristics parameter；
The initial characteristics parameter information in combined radiation source and state plateau, synthesis obtain initial characteristics input quantity.
Step 2: constructing " initial characteristics-neural network middle layer A- radiation source and dress using initial characteristics as input quantity
Carrying platform classification " upper layer Classification Neural structure, the information in the Classification Neural of upper layer is to propagate in one direction, is not had
Reversed information is propagated, the training neural network middle layer A in a manner of supervised learning, eventually by neural network middle layer A
The eigenmatrix A of output mapping initial characteristics and radiation source, loading platform class relations；
Specifically, each neuron according to information is received successively is divided into different groups in Classification Neural, each
Group regards a nervous layer as, and the neuron in each nervous layer receives the output signal of neuron in one layer of nervous layer, and after
The continuous neuron output a signal in next nervous layer；High dimension table of each nervous layer as data information in input signal
Show, is considered as a nonlinear function, by the way that simple non-linear functions are repeatedly compound, realization input signal to output signal
Complex mappings；
Neuron in Classification Neural not only receives the signal of other neurons output, at the same also according to radiation source and
Loading platform classification, which compares, receives oneself feedback signal, the parameter of neural network middle layer A is modified by feedback signal, allows
Interbed preferably extracts characteristic parameter.
Step 3: regarding initial characteristics as input quantity and output quantity simultaneously, " initial characteristics-encoder-neural network are constructed
The sparse self-encoding encoder network structure of middle layer B- decoder " lower layer, in sparse self-encoding encoder network encoder to initial characteristics into
The kernel information of initial characteristics is refined in row dimension-reduction treatment；Decoder trains encoder, and whether the information for judging that encoder refines is quasi-
Really, if obtain the feature with initial characteristics identical information amount, and the error of output is fed back into initial characteristics, mind is trained with this
Through network middle layer B, the eigenmatrix B after initial characteristics are refined by depth is exported.
Specifically, sparse self-encoding encoder is a kind of unsupervised machine learning algorithm, by calculating output and original from coding
The error of input constantly regulate the parameter of self-encoding encoder, finally trains model；Self-encoding encoder can be used for compressing input letter
Breath, extracts useful input feature vector.
Self-encoding encoder is divided into encoder and decoder, and d dimensional feature is transformed into p dimensional feature by encoder, and decoder is by p Wei Te
Sign reconstructs back d dimensional feature, and when meeting p < d, self-encoding encoder is used as dimensionality reduction feature extraction, in addition coding sparsity and value range
Etc. constraint conditions, obtain significant output quantity.
It is as shown in Figure 2: to indicate input without label data, i.e. initial characteristics input quantity, target with { x1, x2, x3, }
It is to export to obtain hW,b(x) ≈ x, that is to say, that sparse self-encoding encoder is to attempt to approach an identity function, so that outputIt is close
In input x；
When the neuronal quantity in neural network middle layer B is less than input quantity, self-encoding encoder neural network is forced to go to learn
" compression " for practising input data indicates；When the neuronal quantity of neural network middle layer B is more, by coding nerve net certainly
Network applies sparsity constraints condition to reach compression input information, extracts the effect of input feature vector.
It is understood that so-called sparsity can be explained are as follows: think its quilt when the output of neuron is close to 1
Activation, and export close to 0 when, thinks that it is suppressed, then the neuron most of the time is made to be in the state of being suppressed
Limitation is referred to as sparsity and limits.
Use bjIndicate the activity of neuron j in neural network middle layer B, bj[x (i)] is indicated in given input x (i)
In the case of, the activity of neuron j, then the average activity of neuron jIt is expressed as follows:
In formula, m indicates the quantity of input x.
The then sparsity parameter ρ of neural network constraint condition:
Wherein, ρ is usually one close to 0 lesser value.
Further, it in order to realize that sparsity limits, needs to limitIt is kept in smaller range with the value of ρ, setting punishment
The factor is as follows:
In formula, S2Indicate the neuronal quantity in neural network middle layer B；
Based on relative entropy, penalty factor is indicated are as follows:
Wherein,Be one using ρ as mean value and one withFor mean value
Two Bernoulli random variables between relative entropy；
WhenWhen,And withDifference between ρ increases and monotonic increase is presented.
Sparsity parameter ρ=0.2 is set,WithVariation it is as shown in Figure 3；It is found that relative entropy existsWhen reach its minimum value 0, and work asWhen close 0 or 1, relative entropy is then become very large, and is intended to ∞, because
This, which minimizes penalty factor, to makeClose to ρ.
To obtain the overall cost function of sparse self-encoding encoder neural network are as follows:
In formula, β indicates the weight of control sparsity penalty factor, this algorithm primary data dimension can be greatly reduced.
Self-encoding encoder arrives decoder to neural network middle layer B by encoder again, between layers full connection mutually.It is logical
Cross minimum reconfiguring false, can effectively learning network parameter, desired characteristic parameter is obtained with this.
Step 4: the eigenmatrix A that will reflect initial characteristics, radiation source and loading platform class relations, initial with reflection
The eigenmatrix B of feature itself inherent attribute is stitched together, and obtains final complex environment characteristic parameter.
Classification Neural identification and sparse self-encoding encoder neural network recognization are wanted the method combined by the present invention, are deeply divided
The essence of analysis and research emitter Signals explores new characteristic parameter, and building is more conducive to the feature vector of signal identification, is promoted
The recognition capability of radar emitter signal under complex environment.
The foregoing is merely illustrative of the preferred embodiments of the present invention, the protection scope being not intended to limit the invention, any
Those skilled in the art within the spirit and principles in the present invention made by any modifications, equivalent replacements, and improvements etc.,
It should all be included in the protection scope of the present invention.
Claims (4)
1. a kind of Radar emitter characteristic parameter extraction method under complex environment based on deep learning characterized by comprising
Initial characteristics extract: extracting the parameter information of radiation source and loading platform as initial characteristics；
Classification Neural building: input initial characteristics construct " initial characteristics-neural network middle layer A- radiation source and loading
Platform classification " upper layer Classification Neural structure exports mapping initial characteristics and radiation source, dress by neural network middle layer A
The eigenmatrix A of carrying platform class relations；
Sparse self-encoding encoder network struction: initial characteristics are used as input and output amount simultaneously, construct " initial characteristics-encoder-nerve
The sparse self-encoding encoder network structure of network middle layer B- decoder " lower layer exports initial characteristics quilt by neural network middle layer B
Inherent attribute eigenmatrix B after depth refinement；
Eigenmatrix splicing: the relationship characteristic matrix A and reflection that reflect initial characteristics, radiation source and loading platform classification is initial
The eigenmatrix B of feature itself inherent attribute is stitched together, and obtains final complex environment characteristic parameter.
2. Radar emitter characteristic parameter extraction side under a kind of complex environment based on deep learning according to claim 1
Method, it is characterised in that: the initial characteristics are repeated for the carrier frequency that the parameter of radiation source includes radar, pulsewidth, angle of arrival, pulse
Frequency, antenna scan period, in conjunction with the pulse arrival time of the signals such as communication and interference, pulse envelope parameter, intra-pulse modulation ginseng
Number, amplitude, frequency spectrum parameter；
The initial characteristics include loading platform movement speed, space position parameter for the parameter of loading platform.
3. Radar emitter characteristic parameter extraction side under a kind of complex environment based on deep learning according to claim 1
Method, it is characterised in that: the information in the Classification Neural is propagated in one direction, the instruction of the neural network middle layer A
The mode of white silk is by the way of supervised learning.
4. Radar emitter characteristic parameter extraction side under a kind of complex environment based on deep learning according to claim 1
Method, it is characterised in that: in the sparse self-encoding encoder network, encoder is used to carry out dimension-reduction treatment to initial characteristics, refines just
The kernel information of beginning feature；For decoder for training encoder, whether the information for judging that encoder refines accurate, if obtain with
The feature of initial characteristics identical information amount, and the error of output is fed back into initial characteristics, neural network middle layer is trained with this
B exports the eigenmatrix B after initial characteristics are refined by depth.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910462165.3A CN110187321B (en) | 2019-05-30 | 2019-05-30 | Radar radiation source characteristic parameter extraction method based on deep learning in complex environment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910462165.3A CN110187321B (en) | 2019-05-30 | 2019-05-30 | Radar radiation source characteristic parameter extraction method based on deep learning in complex environment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110187321A true CN110187321A (en) | 2019-08-30 |
| CN110187321B CN110187321B (en) | 2022-07-22 |
Family
ID=67718895
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910462165.3A Active CN110187321B (en) | 2019-05-30 | 2019-05-30 | Radar radiation source characteristic parameter extraction method based on deep learning in complex environment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110187321B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111630787A (en) * | 2020-04-07 | 2020-09-04 | 东莞理工学院 | MIMO multi-antenna signal transmission and detection technology based on deep learning |
| CN112034434A (en) * | 2020-09-04 | 2020-12-04 | 中国船舶重工集团公司第七二四研究所 | Radar radiation source identification method based on sparse time-frequency detection convolutional neural network |
| CN112308008A (en) * | 2020-11-12 | 2021-02-02 | 电子科技大学 | Radar radiation source individual identification method based on working mode open set of transfer learning |
| CN112859025A (en) * | 2021-01-05 | 2021-05-28 | 河海大学 | Radar signal modulation type classification method based on hybrid network |
| CN117347961A (en) * | 2023-12-04 | 2024-01-05 | 中国电子科技集团公司第二十九研究所 | Radar function attribute identification method based on Bayesian learning |
Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4025920A (en) * | 1972-09-28 | 1977-05-24 | Westinghouse Electric Corporation | Identification of radar systems |
| US4945494A (en) * | 1989-03-02 | 1990-07-31 | Texas Instruments Incorporated | Neural network and system |
| CN105512680A (en) * | 2015-12-02 | 2016-04-20 | 北京航空航天大学 | Multi-view SAR image target recognition method based on depth neural network |
| CN107194433A (en) * | 2017-06-14 | 2017-09-22 | 电子科技大学 | A kind of Radar range profile's target identification method based on depth autoencoder network |
| CN107238822A (en) * | 2017-06-13 | 2017-10-10 | 电子科技大学 | True and false target one-dimensional range profile Nonlinear Orthogonal subspace representation method |
| CN107545903A (en) * | 2017-07-19 | 2018-01-05 | 南京邮电大学 | A kind of phonetics transfer method based on deep learning |
| CN107610692A (en) * | 2017-09-22 | 2018-01-19 | 杭州电子科技大学 | The sound identification method of self-encoding encoder multiple features fusion is stacked based on neutral net |
| CN107832787A (en) * | 2017-10-31 | 2018-03-23 | 杭州电子科技大学 | Recognition Method of Radar Emitters based on bispectrum own coding feature |
| CN108090412A (en) * | 2017-11-17 | 2018-05-29 | 西北工业大学 | A kind of radar emission source category recognition methods based on deep learning |
| WO2018106805A1 (en) * | 2016-12-09 | 2018-06-14 | William Marsh Rice University | Signal recovery via deep convolutional networks |
| EP3425421A1 (en) * | 2017-07-07 | 2019-01-09 | Infineon Technologies AG | System and method for identifying a biological target using radar sensors |
| CN109285168A (en) * | 2018-07-27 | 2019-01-29 | 河海大学 | A kind of SAR image lake boundary extraction method based on deep learning |
| CN109343046A (en) * | 2018-09-19 | 2019-02-15 | 成都理工大学 | Radar gait recognition method based on multifrequency multiple domain deep learning |
| US20190087726A1 (en) * | 2017-08-30 | 2019-03-21 | The Board Of Regents Of The University Of Texas System | Hypercomplex deep learning methods, architectures, and apparatus for multimodal small, medium, and large-scale data representation, analysis, and applications |
| CN109545227A (en) * | 2018-04-28 | 2019-03-29 | 华中师范大学 | Speaker's gender automatic identifying method and system based on depth autoencoder network |
| CN109614905A (en) * | 2018-12-03 | 2019-04-12 | 中国人民解放军空军工程大学 | A kind of radar emitter signal depth intrapulse feature extraction method |
| US20190138860A1 (en) * | 2017-11-08 | 2019-05-09 | Adobe Inc. | Font recognition using adversarial neural network training |
-
2019
- 2019-05-30 CN CN201910462165.3A patent/CN110187321B/en active Active
Patent Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4025920A (en) * | 1972-09-28 | 1977-05-24 | Westinghouse Electric Corporation | Identification of radar systems |
| US4945494A (en) * | 1989-03-02 | 1990-07-31 | Texas Instruments Incorporated | Neural network and system |
| CN105512680A (en) * | 2015-12-02 | 2016-04-20 | 北京航空航天大学 | Multi-view SAR image target recognition method based on depth neural network |
| WO2018106805A1 (en) * | 2016-12-09 | 2018-06-14 | William Marsh Rice University | Signal recovery via deep convolutional networks |
| CN107238822A (en) * | 2017-06-13 | 2017-10-10 | 电子科技大学 | True and false target one-dimensional range profile Nonlinear Orthogonal subspace representation method |
| CN107194433A (en) * | 2017-06-14 | 2017-09-22 | 电子科技大学 | A kind of Radar range profile's target identification method based on depth autoencoder network |
| EP3425421A1 (en) * | 2017-07-07 | 2019-01-09 | Infineon Technologies AG | System and method for identifying a biological target using radar sensors |
| CN107545903A (en) * | 2017-07-19 | 2018-01-05 | 南京邮电大学 | A kind of phonetics transfer method based on deep learning |
| US20190087726A1 (en) * | 2017-08-30 | 2019-03-21 | The Board Of Regents Of The University Of Texas System | Hypercomplex deep learning methods, architectures, and apparatus for multimodal small, medium, and large-scale data representation, analysis, and applications |
| CN107610692A (en) * | 2017-09-22 | 2018-01-19 | 杭州电子科技大学 | The sound identification method of self-encoding encoder multiple features fusion is stacked based on neutral net |
| CN107832787A (en) * | 2017-10-31 | 2018-03-23 | 杭州电子科技大学 | Recognition Method of Radar Emitters based on bispectrum own coding feature |
| US20190138860A1 (en) * | 2017-11-08 | 2019-05-09 | Adobe Inc. | Font recognition using adversarial neural network training |
| CN108090412A (en) * | 2017-11-17 | 2018-05-29 | 西北工业大学 | A kind of radar emission source category recognition methods based on deep learning |
| CN109545227A (en) * | 2018-04-28 | 2019-03-29 | 华中师范大学 | Speaker's gender automatic identifying method and system based on depth autoencoder network |
| CN109285168A (en) * | 2018-07-27 | 2019-01-29 | 河海大学 | A kind of SAR image lake boundary extraction method based on deep learning |
| CN109343046A (en) * | 2018-09-19 | 2019-02-15 | 成都理工大学 | Radar gait recognition method based on multifrequency multiple domain deep learning |
| CN109614905A (en) * | 2018-12-03 | 2019-04-12 | 中国人民解放军空军工程大学 | A kind of radar emitter signal depth intrapulse feature extraction method |
Non-Patent Citations (3)
| Title |
|---|
| 刘家麒 等: "基于注意力机制和双向GRU模型的雷达HRRP目标识别", 《雷达学报》 * |
| 吴晨桐: "基于多域特征提取的雷达辐射源识别", 《中国优秀硕士学位论文全文数据库》 * |
| 阮怀玉: "基于稀疏表示和深度学习的SAR图像目标识别研究", 《中国优秀硕士学位论文全文数据库》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111630787A (en) * | 2020-04-07 | 2020-09-04 | 东莞理工学院 | MIMO multi-antenna signal transmission and detection technology based on deep learning |
| CN112034434A (en) * | 2020-09-04 | 2020-12-04 | 中国船舶重工集团公司第七二四研究所 | Radar radiation source identification method based on sparse time-frequency detection convolutional neural network |
| CN112034434B (en) * | 2020-09-04 | 2022-05-20 | 中国船舶重工集团公司第七二四研究所 | Radar radiation source identification method based on sparse time-frequency detection convolutional neural network |
| CN112308008A (en) * | 2020-11-12 | 2021-02-02 | 电子科技大学 | Radar radiation source individual identification method based on working mode open set of transfer learning |
| CN112859025A (en) * | 2021-01-05 | 2021-05-28 | 河海大学 | Radar signal modulation type classification method based on hybrid network |
| CN112859025B (en) * | 2021-01-05 | 2023-12-01 | 河海大学 | Radar signal modulation type classification method based on hybrid network |
| CN117347961A (en) * | 2023-12-04 | 2024-01-05 | 中国电子科技集团公司第二十九研究所 | Radar function attribute identification method based on Bayesian learning |
| CN117347961B (en) * | 2023-12-04 | 2024-02-13 | 中国电子科技集团公司第二十九研究所 | Radar function attribute identification method based on Bayesian learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110187321B (en) | 2022-07-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110187321A (en) | Radar emitter characteristic parameter extraction method under complex environment based on deep learning | |
| Li et al. | Work modes recognition and boundary identification of MFR pulse sequences with a hierarchical seq2seq LSTM | |
| Johnson et al. | Pulse-coupled neural networks | |
| Wang et al. | Recognition of radar active‐jamming through convolutional neural networks | |
| CN106295694B (en) | A kind of face identification method of iteration weight set of constraints rarefaction representation classification | |
| CN110109060A (en) | A kind of radar emitter signal method for separating and system based on deep learning network | |
| Gu et al. | Image thinning using pulse coupled neural network | |
| CN106503654A (en) | A kind of face emotion identification method based on the sparse autoencoder network of depth | |
| CN109597043A (en) | Radar Signal Recognition method based on quantum particle swarm convolutional neural networks | |
| Ashrapov | Tabular GANs for uneven distribution | |
| CN110378205A (en) | A kind of Complex Radar Radar recognition algorithm based on modified CNN network | |
| CN108108751A (en) | A kind of scene recognition method based on convolution multiple features and depth random forest | |
| KR102318775B1 (en) | Method for Adaptive EEG signal processing using reinforcement learning and System Using the same | |
| CN113050042A (en) | Radar signal modulation type identification method based on improved UNet3+ network | |
| CN111767848A (en) | Radiation source individual identification method based on multi-domain feature fusion | |
| CN107491729B (en) | Handwritten digit recognition method based on cosine similarity activated convolutional neural network | |
| CN108549832A (en) | LPI radar signal sorting technique based on full Connection Neural Network | |
| CN109726653A (en) | Radar Signal Recognition method based on RNN-DenseNet network | |
| CN113156376B (en) | SACNN-based radar radiation source signal identification method | |
| CN112801297B (en) | Machine learning model adversity sample generation method based on conditional variation self-encoder | |
| CN113343814B (en) | Handwritten digital image recognition method based on single-node photon reserve pool calculation | |
| Zhang et al. | Few-shot learning for fine-grained signal modulation recognition based on foreground segmentation | |
| Shin et al. | A closer look at the intervention procedure of concept bottleneck models | |
| Du et al. | Balanced neural architecture search and its application in specific emitter identification | |
| Zhai et al. | Adaptive feature extraction and fine‐grained modulation recognition of multi‐function radar under small sample conditions |
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 |