CN115014404A - High-precision high-speed fiber grating demodulation method based on deep learning - Google Patents
High-precision high-speed fiber grating demodulation method based on deep learning Download PDFInfo
- Publication number
- CN115014404A CN115014404A CN202210497336.8A CN202210497336A CN115014404A CN 115014404 A CN115014404 A CN 115014404A CN 202210497336 A CN202210497336 A CN 202210497336A CN 115014404 A CN115014404 A CN 115014404A
- Authority
- CN
- China
- Prior art keywords
- fiber grating
- deep learning
- data
- demodulation
- precision
- 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
- 239000000835 fiber Substances 0.000 title claims abstract description 57
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000013135 deep learning Methods 0.000 title claims abstract description 24
- 238000001228 spectrum Methods 0.000 claims abstract description 34
- 230000003595 spectral effect Effects 0.000 claims abstract description 31
- 238000005457 optimization Methods 0.000 claims abstract description 23
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 18
- 238000013136 deep learning model Methods 0.000 claims abstract description 16
- 238000007781 pre-processing Methods 0.000 claims abstract description 13
- 238000004088 simulation Methods 0.000 claims abstract description 13
- 238000012549 training Methods 0.000 claims abstract description 13
- 238000013528 artificial neural network Methods 0.000 claims abstract description 9
- 238000012360 testing method Methods 0.000 claims abstract description 7
- 238000000638 solvent extraction Methods 0.000 claims abstract description 4
- 238000013527 convolutional neural network Methods 0.000 claims description 16
- 230000006870 function Effects 0.000 claims description 14
- 238000006243 chemical reaction Methods 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 238000003062 neural network model Methods 0.000 claims description 5
- 239000013307 optical fiber Substances 0.000 claims description 5
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 3
- 230000004913 activation Effects 0.000 claims description 3
- 238000005315 distribution function Methods 0.000 claims description 2
- 210000002569 neuron Anatomy 0.000 claims description 2
- 230000015654 memory Effects 0.000 description 4
- 238000011176 pooling Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D5/00—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable
- G01D5/26—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable characterised by optical transfer means, i.e. using infrared, visible, or ultraviolet light
- G01D5/32—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable characterised by optical transfer means, i.e. using infrared, visible, or ultraviolet light with attenuation or whole or partial obturation of beams of light
- G01D5/34—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable characterised by optical transfer means, i.e. using infrared, visible, or ultraviolet light with attenuation or whole or partial obturation of beams of light the beams of light being detected by photocells
- G01D5/353—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable characterised by optical transfer means, i.e. using infrared, visible, or ultraviolet light with attenuation or whole or partial obturation of beams of light the beams of light being detected by photocells influencing the transmission properties of an optical fibre
-
- 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
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Optical Transform (AREA)
Abstract
The invention discloses a high-precision high-speed fiber grating demodulation method based on deep learning, which comprises the following steps: constructing a simulation data set based on a fiber grating reflection spectrum theoretical model; carrying out data preprocessing on the simulation data set data; building a neural network based on a deep learning framework; dividing the simulation data set into a training set and a test set, and optimizing the deep learning model by adjusting the network structure, the optimization algorithm and the hyper-parameter of the deep learning model; the demodulation equipment receives the spectrum signal reflected by the fiber grating array and outputs the spectrum signal to the upper computer; partitioning the spectral data and preprocessing the data; and inputting the processed spectral data into a deep learning model to obtain the central wavelength of the fiber grating sensor corresponding to the reflection spectrum, thereby realizing high-precision and high-speed demodulation of the spectral data. The invention is based on deep learning, and solves the problems of slow demodulation rate and low demodulation precision of the traditional algorithm of the fiber bragg grating.
Description
Technical Field
The invention belongs to the technical field of optical fiber sensing, and particularly relates to a high-precision and high-speed optical fiber demodulation method based on deep learning.
Background
The fiber grating sensor is widely applied to a plurality of fields of perimeter security protection, civil engineering, power cables, aerospace and the like due to the advantages of small volume, light weight, electromagnetic interference resistance, easiness in building a sensing network and the like. With the gradual expansion of the application range of the optical fiber sensing technology, the requirements on the measurement precision, the stability and the real-time performance of a sensing system are increasingly improved.
The demodulation method of the fiber grating sensor network can be divided into a direct method, a correlation method, a fitting method, a conversion method and an intelligent algorithm according to the principle. The traditional demodulation method is greatly influenced by noise, contradictions between demodulation rate and demodulation precision exist, the direct method can realize faster demodulation but the demodulation precision is too low, and the correlation method, the fitting method, the conversion method and the intelligent algorithm can improve the demodulation precision but the calculation time is too long.
Disclosure of Invention
The invention aims to provide a high-precision high-speed fiber grating demodulation method based on deep learning, which is based on deep learning technology and combines FBG demodulation models built by CNN and LSTM networks to complete the training and optimization of the models and solve the problems of slow demodulation rate and low demodulation precision of the traditional fiber grating algorithm.
In order to solve the technical problems, the technical scheme of the invention is as follows: a high-precision high-speed fiber grating demodulation method based on deep learning comprises the following steps:
constructing a simulation data set based on a fiber grating reflection spectrum theoretical model;
carrying out data preprocessing on the analog data set data, wherein the data preprocessing comprises format conversion and normalization processing;
building a neural network based on a deep learning framework;
dividing the simulation data set into a training set and a test set, and optimizing the deep learning model by adjusting the network structure, the optimization algorithm and the hyper-parameter of the deep learning model;
the demodulation equipment receives the spectrum signal reflected by the fiber grating array and outputs the spectrum signal to the upper computer;
partitioning the spectral data and preprocessing the data;
and inputting the processed spectral data into a deep learning model to obtain the central wavelength of the fiber grating sensor corresponding to the reflection spectrum, thereby realizing high-precision and high-speed demodulation of the spectral data.
The fiber grating reflection spectrum theoretical model is an approximate Gaussian distribution function of a fiber grating reflection spectrum, and is expressed as follows:
in the formula, R (lambda ) B ) Denotes a central wavelength of λ B Optical fiber grating reflection spectrum of (I) peak Representing the magnitude of the reflected spectrum; delta lambda B The 3dB bandwidth of the fiber grating.
Central wavelength lambda of the generated simulated data set B The spectral amplitude I is arranged in the 1540-1544nm interval peak Is 0.6-0.95mw,. DELTA.lambda B 0.2-0.35nm, adding random white gaussian noise with signal to noise ratios of 15, 17, 19, 21, 23 and 25.
Analog data sets were generated with wavelength resolutions of 1pm, 10pm and 20pm, respectively.
A CNN and LSTM network model is built based on a PyTorch framework, data is imported through a DataLoader, and a training set and a test set are divided into a simulation data set according to the ratio of 7: 3.
The optimization algorithm specifically comprises the following steps: applying Adam optimization algorithm and ReLU activation function, setting the number of samples in a training batch as n, and setting the loss function as n label values lambda Bi And predicted value y i Root mean square error between, expressed as:
when the data preprocessing is carried out on the spectral data, the wavelength and the light intensity are normalized to be in the interval of [0,1] by adopting a maximum and minimum normalization method, and the formula is as follows:
subsequently, the GAF algorithm is used to convert the one-dimensional spectral data in the CNN model into two-dimensional image data.
The deep learning model optimization specifically comprises model optimization of the CNN model on the size of the GAF output image, the size of a convolution kernel, the network depth, an optimization function and a loss function; and performing model optimization on the LSTM network from the number of hidden neurons, the number of stacked layers and an optimization function.
Setting a working wavelength interval of the fiber grating sensor according to a 2nm range, receiving spectral data of a fiber grating sensing array by fiber grating modulation and demodulation equipment based on an adjustable laser, and transmitting the spectral data to an upper computer for processing; after receiving the spectrum data, the data are partitioned according to the range of 2nm, and different spectrum intervals can share one neural network model for demodulation.
And constructing a nonlinear regression model from the fiber bragg grating spectral data to the central wavelength, and receiving the spectral data to enable the neural network to output the central wavelength of the spectrum corresponding to the interval.
Compared with the prior art, the invention has the beneficial effects that:
(1) the invention realizes the conversion from one-dimensional spectral data to two-dimensional image data based on the Gram Angular Field (GAF) technology, and realizes the application of a convolution neural network in the fiber bragg grating demodulation;
(2) based on an expansion data set, the method combines a convolutional neural network and a long-time memory neural network in deep learning to realize and optimize a fiber grating demodulation model, solves the contradiction between the demodulation precision and the demodulation rate of the traditional fiber grating demodulation algorithm, and realizes the high-speed demodulation of the fiber grating spectrum, wherein the demodulation test set rate can reach 1ms, and the demodulation frequency can reach 100 Hz;
(3) the fiber grating demodulation model based on the convolutional neural network and the long-term memory neural network has good generalization capability, can still maintain good verification effect even under the conditions of low wavelength resolution and large noise, has demodulation error RMSE (remote measurement error) less than 0.5pm under spectral data with 20pm wavelength resolution, and realizes high-precision demodulation of a fiber grating spectrum.
Drawings
Fig. 1 is a flowchart of a deep learning-based fiber grating demodulation model according to an embodiment of the present invention;
FIG. 2 is a diagram illustrating a structure of a CNN according to an embodiment of the present invention;
fig. 3 is a schematic diagram of an LSTM network structure in the embodiment of the present invention.
Detailed Description
For a better understanding of the technical content of the present invention, the present invention will be described in further detail with reference to the accompanying drawings and examples, it being understood that the embodiments described herein are merely for purposes of illustration and explanation and are not intended to limit the present invention.
(1) Example 1
The invention provides a fiber grating demodulation method based on deep learning, which comprises the following steps:
step 1: building a simulation data set, namely building the simulation data set based on a fiber grating reflection spectrum theoretical model; establishing a simulation data set based on a fiber grating reflection spectrum theoretical model, and setting a central wavelength lambda B In the 1540-1544nm interval, the wavelength resolution is 10pm, and 400 points in the 4nm interval, the input data of 1 sample is (1, 400). Setting spectral amplitude I peak Is 0.6-0.95mw,. DELTA.lambda B 0.2-0.35nm, adding random white Gaussian noise with signal-to-noise ratios of 15, 17, 19, 21, 23 and 25, and converting the corresponding central wavelength lambda to B As a tag value, the number of generated data set samples was 60000.
Step 2: and (3) data preprocessing, namely completing the conversion and normalization processing of the data format of the data set, and normalizing the data into numerical values in a [0,1] interval.
And step 3: the method comprises the steps of realizing a deep learning model, building a convolutional neural network model based on a PyTorch framework, importing data through a DataLoader, and dividing a training set verification set according to a ratio of 7:3 of a simulation data set. Adam optimization algorithm and ReLU activation function are applied. Setting the number of samples in a training batch to be n, and setting the loss function to be n label values lambda Bi And predicted value y i The root mean square error between, can be expressed as:
and 4, step 4: and (4) optimizing the deep learning model, namely dividing the data set into a training set and a testing set, and adjusting the network structure, the optimization algorithm, the hyper-parameters and other contents of the deep learning model to realize the optimization of the deep learning model.
And 5: and the demodulation equipment receives the spectrum signal reflected by the fiber grating array and outputs the spectrum signal to an upper computer.
Step 6: and (3) data preprocessing, namely partitioning the data according to a range of 2nm, and demodulating different spectral intervals by sharing a neural network model to finish format conversion and normalization processing of the data.
And 7: and (3) demodulating the fiber bragg grating, inputting the processed spectral data into a deep learning model to obtain the central wavelength of the fiber bragg grating sensor corresponding to the reflection spectrum, and demodulating the spectral data at high speed and high precision.
(2) Example 2
This embodiment 2 belongs to an embodiment based on a convolutional neural network, and is a specific scheme of the method embodiment of the above embodiment 1, and please refer to the method embodiment 1 for details that are not described in detail in this embodiment.
As shown in fig. 2, the fiber grating wavelength demodulation model based on the convolutional neural network according to the present invention includes the following steps:
step 1: and data preprocessing, namely converting the one-dimensional spectral data into two-dimensional image data through a GAF technology, converting the one-dimensional data from a Cartesian coordinate system to a polar coordinate system through the GAF technology, encoding a time dimension into a geometrical structure of a Gram matrix, and keeping the position information of the data. The size of the output image data by the GAF technique is 56.
Step 2: c1 and C2 are convolutional layers, P1 and P2 are pooling layers, F1 and F2 are all connected layers, and when an image with the size of 56 is input into C1, the convolution kernel of C1 is 3 × 3, and the feature map is 6, and the output feature size is 6 × 54 × 54. Through pooling layer P1, maximum pooling is employed so that the feature map is one-fourth the size of the original. Then, after passing through a round of convolution layer C2 and pooling layer P2, inputting the corresponding characteristic vector, inputting the characteristic vector into full-connection layers F1 and F2, and performing regression to obtain the center wavelength y of the corresponding spectrum i 。
And step 3: and training 300 epochs to obtain corresponding convolutional neural network models.
And 4, step 4: the convolutional neural network model is optimized in terms of convolutional kernel size and network depth, optimization algorithm and loss function.
(3) Example 3
Embodiment 3 belongs to an embodiment based on a convolutional neural network, and is a specific scheme of the method embodiment of embodiment 1, and please refer to method embodiment 1 for details that are not described in detail in this embodiment.
As shown in fig. 3, the fiber grating wavelength demodulation model based on the long-and-short-term memory neural network according to the present invention includes the following steps:
step 1: the number of samples for LSTM model learning is set to be 128, the initial learning rate is 0.001, the number of nodes of a hidden layer is 256, the depth of the model can be increased by stacking the hidden layers of the LSTM, the number of stacked layers of the model is set to be 1, and the hidden layers contain offset values.
Step 2: and training 300 epochs to obtain corresponding long-time and short-time memory neural network models.
And step 3: and optimizing the long-time memory-based neural network model from the aspects of the number of hidden units, an optimization algorithm, the number of stacked layers of hidden layers and a loss function.
Although the present invention has been described with reference to the preferred embodiments, it is not intended to be limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the invention. Therefore, the protection scope of the present invention should be determined by the appended claims.
In the description of the specification, reference to the description of "one embodiment", "preferably", "an example", "a specific example" or "some examples", etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention, and schematic representations of the terms in this specification do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
With the above structure and principle in mind, those skilled in the art should understand that the present invention is not limited to the above embodiments, and modifications and substitutions based on the known technology in the field are within the scope of the present invention, which should be limited by the claims.
Claims (10)
1. A high-precision high-speed fiber grating demodulation method based on deep learning is characterized by comprising the following steps:
constructing a simulation data set based on a fiber grating reflection spectrum theoretical model;
carrying out data preprocessing on the analog data set data, wherein the data preprocessing comprises format conversion and normalization processing;
building a neural network based on a deep learning framework;
dividing the simulation data set into a training set and a test set, and optimizing the deep learning model by adjusting the network structure, the optimization algorithm and the hyper-parameter of the deep learning model;
the demodulation equipment receives the spectrum signal reflected by the fiber grating array and outputs the spectrum signal to the upper computer;
partitioning the spectral data and preprocessing the data;
and inputting the processed spectral data into a deep learning model to obtain the central wavelength of the fiber grating sensor corresponding to the reflection spectrum, thereby realizing high-precision and high-speed demodulation of the spectral data.
2. The demodulation method of the high-precision and high-speed fiber grating based on the deep learning as claimed in claim 1, wherein the theoretical model of the fiber grating reflection spectrum is an approximate gaussian distribution function of the fiber grating reflection spectrum, and is expressed as:
in the formula, R (lambda ) B ) Denotes a central wavelength of λ B Optical fiber grating reflection spectrum of (I) peak Representing the magnitude of the reflected spectrum; delta lambda B The 3dB bandwidth of the fiber grating.
3. The deep learning-based high-precision high-speed fiber grating demodulation method as claimed in claim 2, wherein the center wavelength λ of the spectral data in the generated analog data set B Set in 1540-1544nm interval, spectral amplitude I peak Is 0.6-0.95mw,. DELTA.lambda B 0.2-0.35nm, adding random white gaussian noise with signal to noise ratios of 15, 17, 19, 21, 23 and 25.
4. The method for demodulating the high-precision and high-speed fiber grating based on the deep learning as claimed in claim 1, wherein the analog data sets with the wavelength resolution of 1pm, 10pm and 20pm are respectively generated.
5. The high-precision high-speed fiber grating demodulation method based on deep learning of claim 1, wherein a CNN and LSTM network model is built based on a PyTorch framework, data is imported through a DataLoader, and a training set and a test set are divided into a simulation data set according to the ratio of 7: 3.
6. The high-precision high-speed fiber grating demodulation method based on deep learning as claimed in claim 2, wherein the optimization algorithm specifically comprises: applying Adam optimization algorithm and ReLU activation function, setting the number of samples in a training batch as n, and setting the loss function as n label values lambda Bi And predicted value y i Root mean square error between, expressed as:
7. the method for demodulating high-precision and high-speed fiber bragg grating based on deep learning as claimed in claim 5, wherein when the data preprocessing is performed on the spectral data, the wavelength and the light intensity are normalized to the interval of [0,1] by adopting a maximum and minimum normalization method, and the formula is as follows:
subsequently, the GAF algorithm is used to convert the one-dimensional spectral data in the CNN model into two-dimensional image data.
8. The deep learning-based high-precision high-speed fiber grating demodulation method according to claim 7, wherein the optimization of the deep learning model specifically comprises model optimization of the CNN model on the GAF output image size, the convolution kernel size, the network depth, an optimization function and a loss function; and performing model optimization on the LSTM network from the number of hidden neurons, the number of stacked layers and an optimization function.
9. The high-precision high-speed fiber grating demodulation method based on deep learning as claimed in claim 1, wherein the fiber grating sensor is arranged in an operating wavelength range according to a 2nm range, and the fiber grating modulation and demodulation device based on the tunable laser receives the spectral data of the fiber grating sensor array and transmits the spectral data to the upper computer for processing; after receiving the spectrum data, the data are partitioned according to the range of 2nm, and different spectrum intervals can share one neural network model for demodulation.
10. The deep learning-based high-precision high-speed fiber grating demodulation method as claimed in claim 1, wherein a nonlinear regression model from fiber grating spectral data to a central wavelength is constructed, and the spectral data is received, so that the neural network outputs the central wavelength of the spectrum corresponding to the interval.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210497336.8A CN115014404A (en) | 2022-05-09 | 2022-05-09 | High-precision high-speed fiber grating demodulation method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210497336.8A CN115014404A (en) | 2022-05-09 | 2022-05-09 | High-precision high-speed fiber grating demodulation method based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115014404A true CN115014404A (en) | 2022-09-06 |
Family
ID=83069194
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210497336.8A Pending CN115014404A (en) | 2022-05-09 | 2022-05-09 | High-precision high-speed fiber grating demodulation method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115014404A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116861167A (en) * | 2023-06-12 | 2023-10-10 | 河北工程大学 | FBG spectrum cyclic denoising method based on deep learning |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109282837A (en) * | 2018-10-24 | 2019-01-29 | 福州大学 | Bragg grating based on LSTM network interlocks the demodulation method of spectrum |
| US20210304002A1 (en) * | 2020-03-27 | 2021-09-30 | Battelle Memorial Institute | Data handling and machine learning |
| CN113834509A (en) * | 2021-08-27 | 2021-12-24 | 中国计量大学 | Weak reflection fiber grating string demodulator based on deep learning |
| CN114257313A (en) * | 2021-12-18 | 2022-03-29 | 江苏中天科技股份有限公司 | Fiber grating sensing system noise reduction method based on deep learning and related equipment thereof |
| CN114419392A (en) * | 2022-01-19 | 2022-04-29 | 北京理工大学重庆创新中心 | Hyperspectral snapshot image recovery method, device, equipment and medium |
-
2022
- 2022-05-09 CN CN202210497336.8A patent/CN115014404A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109282837A (en) * | 2018-10-24 | 2019-01-29 | 福州大学 | Bragg grating based on LSTM network interlocks the demodulation method of spectrum |
| US20210304002A1 (en) * | 2020-03-27 | 2021-09-30 | Battelle Memorial Institute | Data handling and machine learning |
| CN113834509A (en) * | 2021-08-27 | 2021-12-24 | 中国计量大学 | Weak reflection fiber grating string demodulator based on deep learning |
| CN114257313A (en) * | 2021-12-18 | 2022-03-29 | 江苏中天科技股份有限公司 | Fiber grating sensing system noise reduction method based on deep learning and related equipment thereof |
| CN114419392A (en) * | 2022-01-19 | 2022-04-29 | 北京理工大学重庆创新中心 | Hyperspectral snapshot image recovery method, device, equipment and medium |
Non-Patent Citations (1)
| Title |
|---|
| 喻国威 等: "基于GADF变换和多尺度CNN的哈密瓜表面农药残留可见－近红外光谱判别方法" * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116861167A (en) * | 2023-06-12 | 2023-10-10 | 河北工程大学 | FBG spectrum cyclic denoising method based on deep learning |
| CN116861167B (en) * | 2023-06-12 | 2023-12-15 | 河北工程大学 | FBG spectrum cyclic denoising method based on deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110365612B (en) | Deep learning beam domain channel estimation method based on approximate message transfer algorithm | |
| Cong et al. | CRB weighted source localization method based on deep neural networks in multi-UAV network | |
| CN108732926A (en) | Networked system method for estimating state based on insufficient information | |
| CN110888105B (en) | DOA estimation method based on convolutional neural network and received signal strength | |
| CN115014404A (en) | High-precision high-speed fiber grating demodulation method based on deep learning | |
| Yang et al. | One-dimensional deep attention convolution network (ODACN) for signals classification | |
| US7603004B2 (en) | Neural network demodulation for an optical sensor | |
| CN110954860A (en) | DOA and polarization parameter estimation method | |
| CN108631817B (en) | Method for predicting frequency hopping signal frequency band based on time-frequency analysis and radial neural network | |
| CN114510960A (en) | Method for recognizing distributed optical fiber sensor system mode | |
| CN114257313A (en) | Fiber grating sensing system noise reduction method based on deep learning and related equipment thereof | |
| CN111586151B (en) | Intelligent city data sharing system and method based on block chain | |
| Xia et al. | Radar waveform recognition based on a two‐stream convolutional network and software defined radio | |
| Butcher et al. | Extending reservoir computing with random static projections: a hybrid between extreme learning and RC | |
| Zang et al. | Optoelectronic convolutional neural networks based on time-stretch method | |
| CN115632970A (en) | Method, device and storage medium for estimating communication interference signal bandwidth under non-Gaussian noise | |
| CN115471428A (en) | Co-phase error correction method based on all-optical diffraction neural network | |
| CN115494466A (en) | Self-calibration method for distributed radar | |
| CN114255392A (en) | Nitrogen dioxide concentration prediction system based on satellite hyperspectral remote sensing and artificial intelligence | |
| Baietto et al. | Lean neural networks for real-time embedded spectral notching waveform design | |
| CN113472404B (en) | Method and device for optimizing digital domain beam forming based on condition generation countermeasure network | |
| CN117580015A (en) | Vehicle sensing method and system in ISAC enabled Internet of vehicles environment | |
| CN105301961A (en) | Large-diameter turntable system model identification method based on RBF neural network | |
| Wang et al. | Stochastic configuration network‐based SAR image target classification approach | |
| CN116384232A (en) | Atmospheric turbulence compensation method based on self-supervision learning under different weather conditions aiming at orbital angular momentum |
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 |