CN115598980A - Self-adaptive optical wavefront prediction and feedforward correction method based on deep learning - Google Patents
Self-adaptive optical wavefront prediction and feedforward correction method based on deep learning Download PDFInfo
- Publication number
- CN115598980A CN115598980A CN202211311663.6A CN202211311663A CN115598980A CN 115598980 A CN115598980 A CN 115598980A CN 202211311663 A CN202211311663 A CN 202211311663A CN 115598980 A CN115598980 A CN 115598980A
- Authority
- CN
- China
- Prior art keywords
- wavefront
- prediction
- correction
- model
- deep learning
- 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
- 238000012937 correction Methods 0.000 title claims abstract description 63
- 238000000034 method Methods 0.000 title claims abstract description 30
- 230000003287 optical effect Effects 0.000 title claims abstract description 16
- 238000013135 deep learning Methods 0.000 title claims abstract description 14
- 238000004088 simulation Methods 0.000 claims abstract description 16
- 238000013528 artificial neural network Methods 0.000 claims abstract description 14
- 238000004458 analytical method Methods 0.000 claims abstract description 10
- 230000004075 alteration Effects 0.000 claims abstract description 9
- 230000003044 adaptive effect Effects 0.000 claims description 13
- 238000012360 testing method Methods 0.000 claims description 11
- 238000012549 training Methods 0.000 claims description 10
- 230000000694 effects Effects 0.000 claims description 8
- 238000011156 evaluation Methods 0.000 claims description 4
- 230000006870 function Effects 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 4
- 230000008014 freezing Effects 0.000 claims description 3
- 238000007710 freezing Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 2
- 238000001514 detection method Methods 0.000 abstract description 2
- 238000003062 neural network model Methods 0.000 abstract description 2
- 230000008901 benefit Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 241001522296 Erithacus rubecula Species 0.000 description 1
- 235000019687 Lamb Nutrition 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000001427 coherent effect Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000005693 optoelectronics Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J9/00—Measuring optical phase difference; Determining degree of coherence; Measuring optical wavelength
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J9/00—Measuring optical phase difference; Determining degree of coherence; Measuring optical wavelength
- G01J2009/002—Wavefront phase distribution
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Automation & Control Theory (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Telescopes (AREA)
- Optical Elements Other Than Lenses (AREA)
Abstract
The invention discloses a self-adaptive optical wavefront prediction and feedforward correction method based on deep learning. Firstly, an atmospheric turbulence distorted wavefront with different time and space frequencies can be simulated and generated by utilizing a set of aberration simulation system, so that a distorted light beam firstly carries out wavefront detection through a Hartmann wavefront sensor 1, then carries out forward prediction of the wavefront through a neural network model carried by a computer, finally converts an output predicted wavefront into a control voltage to be loaded on a deformable mirror, carries out correction performance analysis through a Hartmann wavefront sensor 2, and simultaneously realizes the wavefront prediction and feedforward correction of AO. The method adopts a mode of neural network prediction and feedforward correction, utilizes the distorted wavefront detected in real time to develop forward prediction, and converts the output predicted wavefront into the correction surface shape of the deformable mirror, so that the inherent control delay problem of an AO system can be effectively solved; by adopting a feedforward correction mode, a network prediction model can be effectively simplified, and the deployment and application in an actual system are facilitated.
Description
Technical Field
The invention belongs to the technical field of wavefront prediction and correction control, and relates to a self-adaptive optics wavefront prediction and feedforward correction method based on deep learning, which is suitable for a wavefront aberration correction task based on self-adaptive optics.
Background
The AO technology is an effective means for compensating atmospheric turbulence in real time, however, the actual AO system usually has a time delay of 2-3 sampling periods due to the delay of the readout data of the wavefront sensor, the delay of the calculation of the processor control, and the like. In the case of correcting atmospheric turbulence distorted wavefronts with high temporal frequencies, such time delay errors will cause the compensating wavefront on the anamorphic mirror to lag significantly behind the variations of the distorted wavefront, severely limiting the correction performance of the AO technique.
Prediction techniques can effectively compensate for this delay error, especially methods based on neural network classes have shown greater advantages (see Guo Y, zhong L, min L, et al. Adaptive optics based on machine learning: a review [ J ]. Opto-Electronic Advances, 2022. The open-loop control voltage of the deformable mirror is predicted by using a two-layer Back Propagation (BP) neural network prediction algorithm in the case of SZ, the correction effect of the deformable mirror is improved by about 30 times compared with that of the RLS algorithm in the same wind speed (see SZ, von Yong, chengyo, and the like. Adaptive optics system deformable mirror control voltage prediction [ J ]. Strong laser and particle beams, 2012,24 (6): 1281-1286.). Chen Ying et al used the LSTM network to predict voltages with predicted residuals reduced by about 8 times over BP networks (see Chen Y. Volts prediction based on LSTM prediction neural network [ J ]. Optik.2020,220 (16): 4869-4880.). Meanwhile, chenying et al utilizes the LSTM network to predict the Zernike coefficient, and compared with the method of directly predicting the DM voltage, the method can effectively shorten the training calculation time and has practical value (see Chen Y. LSTM, recurrent neural network prediction on Zernike model coefficients [ J ]. Optik,2020, 203. Chen et al, applied to predict AO system aberrated wavefronts and track fast moving targets, such as low earth orbiting satellites (LEO), using a U-Net based convolutional neural network architecture, have experimentally shown that a fine-tuned neural network can reduce wavefront errors by about 50% over non-predictive methods (see j.g. chen, v.sham, and l.l.liu. "Performance of a U-Net-based neural network for predictive adaptive optics," t.let.2021, 46 (10): 2513-2516.).
However, in the development of wavefront prediction based on the neural network method at present, most researchers aim at directly training and predicting various kinds of continuous distorted wavefront information, and the negative feedback closed-loop correction mode based on the wavefront residual error in the traditional AO is difficult to meet. In 2021, robin Swanson et al simulated slope prediction and closed-loop correction using pseudo-open-loop data, which added generation countermeasure network in combination with LSTM network and CNN network based on dense connection for supervised discrimination respectively (see Swanson R, lamb M, coreia CM, et al. Closed loop prediction control of adaptive optics systems with a coherent neural network [ J ] free notes of the radial adaptive Society,2021,503 (2): 2944-2954.), although both composite networks had some improvement in correction performance compared to the classical neural network prediction method directly applied to closed-loop correction, the complexity of network design also limited the implementation of AO system real-time correction to a large extent.
Therefore, the invention provides an adaptive optics wavefront prediction and feedforward correction method based on deep learning.
Disclosure of Invention
The technical problem solved by the invention is as follows: the problem that the compensation wavefront on the deformable mirror obviously lags behind the distorted wavefront change due to the inherent time delay error of an AO system under the condition of correcting the atmospheric turbulence distorted wavefront with high time frequency is solved.
The technical scheme adopted by the invention is as follows: a self-adaptive optical wavefront prediction and feedforward correction method based on deep learning is realized by the following steps:
step S1: obtaining atmospheric turbulence distortion wavefront through Fourier series method time evolution simulation based on Kolmogorov turbulence statistical theory, HV-57 refractive index structure constant model and Buffton wind speed model according to atmospheric freezing flow hypothesis;
step S2: loading the atmospheric turbulence distortion wavefront with fixed time frequency generated by simulation into an aberration simulation system, and acquiring wavefront data by adopting a Hartmann wavefront sensor 1 based on a fixed sampling frame frequency (1 kHz);
and step S3: preprocessing the wavefront data acquired in the step S2, dividing the processed data into a training set and a test set according to the requirements of a prediction model, and performing iterative retraining on the training set data through a pre-established neural network structure to enable the network to adapt to the reconstructed wavefront characteristic information of the Hartmann wavefront sensor 1;
and step S4: performing prediction model test by using the test set data, performing residual error analysis on the output prediction wavefront and the actual label wavefront to be corrected, selecting an optimal model according to the residual error analysis result, and deploying the optimal model in the controller;
step S5: converting the predicted wavefront output by the network model into control voltage to be loaded on a deformable mirror to generate a correction surface shape, and observing a correction residual wavefront through a Hartmann wavefront sensor 2 for evaluating correction performance;
step S6: simulating and generating atmospheric turbulence distortion wavefront with gradually increasing time frequency, loading the atmospheric turbulence distortion wavefront into an aberration simulation system, carrying out wavefront prediction and feedforward correction, and obtaining correction performance under atmospheric turbulence of different time frequencies;
further, in step S1, kolmogorov turbulence statistics theory is that the atmospheric refractive index structural function in the inertia region satisfies "two-thirds law".
Further, in step S3, the data in the training set and the test set need to satisfy the correspondence between the sample and the label, so that the network can correctly learn the mapping relationship between the historical wavefront and the predicted wavefront, the evaluation index of the residual error analysis is the Root Mean Square (RMS) value of the residual wavefront, and the model with the minimum RMS value is selected as the optimal model structure.
Further, the evaluation index of the correction performance in step S5 is a root mean square value (RMS) of the correction residual wavefront detected by the hartmann wavefront sensor 2, and the smaller the RMS value, the better the correction effect is proved.
Further, the feedforward correction in step S6 means that the hartmann wavefront sensor is placed in front of the deformable mirror, so that a distorted wavefront is detected instead of a residual wavefront, and the distorted wavefront can be corrected by converting wavefront information directly predicted by the prediction model into a shape-corrected surface of the deformable mirror.
The principle of the invention is as follows: the invention provides a self-adaptive optical wavefront prediction and feedforward correction method based on deep learning, which is characterized in that atmospheric turbulence distorted wavefronts with different time and space frequencies are generated by simulation, so that distorted light beams are subjected to wavefront detection through a Hartmann wavefront sensor 1, wavefront prediction is performed through a neural network model carried by a computer, output predicted wavefronts are converted into control voltages and loaded onto a deformable mirror, corrected residual wavefronts are observed through a Hartmann wavefront sensor 2, correction performance analysis is performed, and the wavefront prediction and feedforward correction of AO are realized at the same time.
Compared with the prior art, the invention has the following advantages:
1. the method adopts a mode of neural network prediction and feedforward correction, utilizes the distorted wavefront detected in real time to develop forward prediction, and converts the output predicted wavefront into the correction surface shape of the deformable mirror, thereby effectively overcoming the inherent control delay problem of the AO system.
2. The invention adopts a feedforward correction mode, is different from the traditional negative feedback closed-loop correction mode based on wavefront residual errors, can effectively simplify a network prediction model, and is more beneficial to deployment and application in an actual system.
Drawings
FIG. 1 is a flowchart of the working process of the adaptive optical wavefront prediction and feedforward correction method based on deep learning according to the present invention;
FIG. 2 is a diagram of the experimental principle of feedforward correction and the layout of optical path.
Detailed Description
In order to make the technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings in conjunction with specific embodiments.
Fig. 1 is a work flow chart of an adaptive optics wavefront prediction and feedforward correction method based on deep learning, and the specific steps are as follows:
step S1: obtaining atmospheric turbulence distortion wavefront through Fourier series method time evolution simulation based on Kolmogorov turbulence statistical theory, HV-57 refractive index structure constant model and Buffton wind speed model according to atmospheric freezing flow hypothesis;
step S2: loading the atmospheric turbulence distortion wavefront with fixed time frequency generated by simulation into an aberration simulation system, and acquiring wavefront data by adopting a Hartmann wavefront sensor 1 based on a fixed sampling frame frequency (1 kHz);
and step S3: preprocessing the wavefront data acquired in the step S2, dividing the processed data into a training set and a test set according to the requirements of a prediction model, and performing iterative retraining on the training set data through a pre-established neural network structure to enable the network to adapt to the reconstructed wavefront characteristic information of the Hartmann wavefront sensor 1;
and step S4: performing prediction model test by using the test set data, performing residual error analysis on the output prediction wavefront and the actual label wavefront to be corrected, selecting an optimal model according to the residual error analysis result, and deploying the optimal model in the controller;
step S5: converting the predicted wavefront output by the network model into control voltage to be loaded on a deformable mirror to generate a correction surface shape, and observing a correction residual wavefront through a Hartmann wavefront sensor 2 for evaluating correction performance;
step S6: simulating and generating atmospheric turbulence distortion wavefront with gradually increasing time frequency, loading the atmospheric turbulence distortion wavefront into an aberration simulation system, carrying out wavefront prediction and feedforward correction, and obtaining correction performance under atmospheric turbulence of different time frequencies;
in step S1, the Kolmogorov turbulence statistical theory is that the atmospheric refractive index structural function in the inertia region satisfies the "two-thirds law".
Refractive index structure function D n (r) satisfying "two-thirds of the law" can be expressed as:
wherein r is a scalar distance,
for the refractive index structure constant at altitude h, it is commonly expressed by the model HV-57:
wherein, generally, the following are taken: v =21m/s, A =1.7 × 10 -14 。
However, in laser atmosphere transmission and adaptive optical correction techniques, the atmospheric coherence length r is widely adopted 0 To describe the effects of turbulence effects and to evaluate the effect of laser transmission and its phase correction. r is 0 The larger the value, the better the atmospheric conditions, and generally, r 0 And
the relationship of (c) is:
wherein k is a wave number, L is a transmission distance,
is a zenith angle.
When the laser is transmitted at atmospheric level, i.e.
r 0 Expressed as:
the Buffton wind speed model v (h) is expressed as the variation of wind speed in the atmosphere with altitude:
wherein, generally, v is taken g ＝5m/s，v t ＝30m/s，h pk ＝9400，h scale ＝4800。
The RMS value calculation formula for the residual wavefront in steps S4 and S5 is:
where N is the pixel resolution of the distorted wavefront, i.e., the pixel point value, x jpre Is a predicted value, x, of a distorted wavefront pixel point, j jtrue Is the true value, x, of a distorted wavefront pixel point j i The average value of all pixel points in the distorted wavefront is obtained.
FIG. 2 shows the experimental principle of feedforward correction and the layout of optical path in the present invention; in the experiment, the relation between the position of each optical component and the optical path is indicated by arrows in the figure, wherein an aberration simulation system is used for simulating and generating continuous distorted wavefront under different atmospheric turbulence conditions, a Hartmann wavefront sensor 1 is used for detecting the wavefront information, a controller with a prediction model is used for performing wavefront prediction processing to output predicted wavefront, a deformable mirror is used for correcting the predicted wavefront, and a Hartmann wavefront sensor 2 is used for observing residual wavefront to analyze the correction effect.
The present invention is not intended to be limited to the particular embodiments described above, which are intended to be illustrative only and not limiting. Those skilled in the art, having the benefit of this disclosure, may effect numerous modifications thereto and changes may be made without departing from the scope of the invention as defined by the appended claims. The invention has not been described in detail and is part of the common general knowledge of a person skilled in the art.
Claims (5)
1. A self-adaptive optical wavefront prediction and feedforward correction method based on deep learning is characterized in that: the method comprises the following steps:
step S1: according to the assumption of atmospheric freezing flow, obtaining atmospheric turbulence distortion wavefront through Fourier series method time evolution simulation based on Kolmogorov turbulence statistical theory, HV-57 refractive index structural constant model and Buffton wind speed model;
step S2: loading the atmospheric turbulence distortion wavefront with fixed time frequency generated by simulation into an aberration simulation system, and acquiring wavefront data by adopting a Hartmann wavefront sensor 1 based on a fixed sampling frame frequency;
and step S3: preprocessing the wavefront data acquired in the step S2, dividing the processed data into a training set and a test set according to the requirements of a prediction model, and performing iterative retraining on the training set data through a pre-established neural network structure to enable the network to adapt to the reconstructed wavefront characteristic information of the Hartmann wavefront sensor 1;
and step S4: performing prediction model test by using the test set data, performing residual error analysis on the output prediction wavefront and the actual label wavefront to be corrected, selecting an optimal model according to the residual error analysis result, and deploying the optimal model in the controller;
step S5: converting the predicted wavefront output by the network model into control voltage to be loaded on a deformable mirror to generate a correction surface shape, and observing a correction residual wavefront through a Hartmann wavefront sensor 2 for evaluating correction performance;
step S6: and (3) loading the atmospheric turbulence distorted wavefront with gradually increasing simulation generation time frequency into an aberration simulation system, carrying out wavefront prediction and feedforward correction, and obtaining the correction performance under atmospheric turbulence of different time frequencies.
2. The method of claim 1 for adaptive optical wavefront prediction and feed-forward correction based on deep learning, wherein: in the step S1, the Kolmogorov turbulence statistical theory is that the atmospheric refractive index structure function in the inertial region satisfies the two-thirds law.
3. The method of claim 1 for adaptive optical wavefront prediction and feed-forward correction based on deep learning, wherein: the data in the training set and the test set in the step S3 should meet the requirement that the sample corresponds to the label, so that the network can correctly learn the mapping relationship between the historical wavefront and the predicted wavefront, the evaluation index of the residual error analysis is the Root Mean Square (RMS) value of the residual wavefront, and the model with the minimum RMS value is selected as the optimal model structure.
4. The method of claim 1 for adaptive optical wavefront prediction and feed-forward correction based on deep learning, wherein: the evaluation index of the correction performance in step S5 is a root mean square value (RMS) of the correction residual wavefront detected by the hartmann wavefront sensor 2, and the smaller the RMS value, the better the correction effect is proved.
5. The method of claim 1 for adaptive optical wavefront prediction and feed-forward correction based on deep learning, wherein: the feedforward correction in the step S6 means that the Hartmann wavefront sensor is placed in front of the deformable mirror, so that the distorted wavefront is detected instead of the residual wavefront, and the distorted wavefront can be corrected by converting the wavefront information directly predicted by the prediction model into the correction surface shape of the deformable mirror.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211311663.6A CN115598980A (en) | 2022-10-25 | 2022-10-25 | Self-adaptive optical wavefront prediction and feedforward correction method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211311663.6A CN115598980A (en) | 2022-10-25 | 2022-10-25 | Self-adaptive optical wavefront prediction and feedforward correction method based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115598980A true CN115598980A (en) | 2023-01-13 |
Family
ID=84848169
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211311663.6A Pending CN115598980A (en) | 2022-10-25 | 2022-10-25 | Self-adaptive optical wavefront prediction and feedforward correction method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115598980A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113310902A (en) * | 2021-05-26 | 2021-08-27 | 中国科学院光电技术研究所 | Optical cavity ring-down adaptive optical active transverse mode matching method |
-
2022
- 2022-10-25 CN CN202211311663.6A patent/CN115598980A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113310902A (en) * | 2021-05-26 | 2021-08-27 | 中国科学院光电技术研究所 | Optical cavity ring-down adaptive optical active transverse mode matching method |
| CN113310902B (en) * | 2021-05-26 | 2023-10-03 | 中国科学院光电技术研究所 | Cavity ring-down self-adaptive optical active transverse mode matching method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11409347B2 (en) | Method, system and storage medium for predicting power load probability density based on deep learning | |
| US5649065A (en) | Optimal filtering by neural networks with range extenders and/or reducers | |
| CN109343505A (en) | Gear method for predicting residual useful life based on shot and long term memory network | |
| CN114331233B (en) | Method and device for estimating total primary productivity of vegetation, electronic equipment and storage medium | |
| CN114004342B (en) | Laser communication system distortion wavefront prediction method based on LSTM network | |
| CN111144644B (en) | Short-term wind speed prediction method based on variation variance Gaussian process regression | |
| CN110991721A (en) | Short-term wind speed prediction method based on improved empirical mode decomposition and support vector machine | |
| CN115598980A (en) | Self-adaptive optical wavefront prediction and feedforward correction method based on deep learning | |
| CN111221123B (en) | Wavefront-sensor-free self-adaptive optical correction method based on model | |
| CN114966685A (en) | Dam deformation monitoring and predicting method based on InSAR and deep learning | |
| CN114970952B (en) | Photovoltaic output short-term prediction method and system considering environmental factors | |
| CN113225130B (en) | Atmospheric turbulence equivalent phase screen prediction method based on machine learning | |
| CN115660887A (en) | Photovoltaic output prediction method and system based on limited weather forecast information | |
| CN111968047A (en) | Adaptive optical image blind restoration method based on generating type countermeasure network | |
| CN116542383A (en) | Distributed photovoltaic system output prediction method based on small fluctuation weather satellite cloud image | |
| CN116780511A (en) | SARIMA model-based power system inertia prediction method | |
| CN114528638A (en) | Ship motion multi-step real-time prediction mixing method and system based on reinforcement learning | |
| Archinuk et al. | Mitigating the nonlinearities in a pyramid wavefront sensor | |
| CN112462600A (en) | High-energy laser control method and system, electronic equipment and storage medium | |
| CN111967660A (en) | Ultra-short-term photovoltaic prediction residual error correction method based on SVR | |
| CN116681154A (en) | Photovoltaic power calculation method based on EMD-AO-DELM | |
| Zhou et al. | Tidal forecasting based on ARIMA-LSTM neural network | |
| CN112036672B (en) | New energy power generation ultra-short term power prediction method and system based on iterative correction | |
| CN115078305A (en) | Zernike coefficient wavefront prediction algorithm based on graph neural network | |
| CN114994800A (en) | Inversion method and device for atmospheric fine particles |
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 |