CN113654670B - Photon-lacking aperture centroid displacement estimation method based on neural network - Google Patents
Photon-lacking aperture centroid displacement estimation method based on neural network Download PDFInfo
- Publication number
- CN113654670B CN113654670B CN202110933080.6A CN202110933080A CN113654670B CN 113654670 B CN113654670 B CN 113654670B CN 202110933080 A CN202110933080 A CN 202110933080A CN 113654670 B CN113654670 B CN 113654670B
- Authority
- CN
- China
- Prior art keywords
- aperture
- neural network
- centroid
- sub
- displacement
- 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.)
- Active
Links
- 238000006073 displacement reaction Methods 0.000 title claims abstract description 72
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 44
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000012549 training Methods 0.000 claims description 23
- 239000011159 matrix material Substances 0.000 claims description 17
- 238000010276 construction Methods 0.000 claims description 2
- 238000013527 convolutional neural network Methods 0.000 claims description 2
- 238000011084 recovery Methods 0.000 abstract description 5
- 238000001514 detection method Methods 0.000 abstract description 4
- 238000005259 measurement Methods 0.000 abstract description 3
- 238000009828 non-uniform distribution Methods 0.000 abstract 2
- 238000012360 testing method Methods 0.000 description 12
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000005484 gravity Effects 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Images
Classifications
-
- 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
- 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
-
- 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
-
- 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
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (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)
- Spectroscopy & Molecular Physics (AREA)
- Testing Of Optical Devices Or Fibers (AREA)
Abstract
The invention discloses a photon-lacking aperture centroid displacement estimation method based on a neural network, which establishes a nonlinear relation between the local part of a shack-Hartmann wavefront sensor and the full-aperture sub-aperture centroid displacement through the neural network, estimates the photon-lacking aperture centroid displacement according to the detectable sub-aperture centroid displacement, can effectively reduce the influence of the sub-aperture light shortage on the detection precision of the wavefront sensor and the robustness of a wavefront recovery algorithm, improves the adaptability of the sensor to the non-uniform distribution of the near-field intensity of a light beam, enhances the wavefront detection stability under the condition that the sub-aperture has the light shortage, and is expected to be used for high-precision and high-robustness wavefront measurement under the conditions of non-uniform distribution of the near-field light intensity, light intensity flicker and the like.
Description
Technical Field
The invention belongs to the technical field of wavefront sensing, and particularly relates to a photon-lacking aperture centroid displacement estimation method based on a neural network.
Background
The shack-Hartmann wavefront sensor (SH-WFS) has the advantages of simple structure, high measuring speed, strong adaptability and the like, and is widely applied to various fields of laser transmission, astronomical observation, medical imaging and the like. The sensor consists of a micro lens array and a CCD. The incident beam is divided, sampled and focused on the CCD through the micro lens array to form a light spot array image. The sub-spot centroid displacement is proportional to the average slope of the corresponding sub-aperture wavefront, the local wavefront slope is estimated by calculating the centroid displacement of each sub-spot, and finally the whole incident wavefront is reconstructed by a corresponding wavefront reconstruction algorithm according to the sub-aperture slope information. Therefore, the accuracy and the completeness of the sub-aperture light spot centroid displacement data directly determine the measurement accuracy of the shack-Hartmann wavefront sensor.
In practical application, under the influence of factors such as atmospheric turbulence intensity and beacon light shortage characteristics, when a light beam reaches a shack-Hartmann wavefront sensor, the intensity of partial sub-aperture light spots is weakened or even disappears in noise, so that the sensor cannot obtain complete and accurate sub-aperture centroid displacement information, when the number of the light-lacking apertures is large, the wavefront restoration precision and the robustness of a restoration algorithm are seriously influenced, and the correction effect of a self-adaptive optical system is limited.
For the problem of partial sub-aperture light shortage, the most common method is the photon-lacking aperture slope nulling method (see weiping, lisneryang, raxi, and lixiang, "influence of partial sub-aperture light shortage on wavefront restoration of shack-hartmann wavefront sensor". china laser 47(4), 0409002,2020). The method improves the wave front restoration precision by discarding inaccurate sub-aperture slope information, and can obtain a wave front restoration effect with higher precision under the condition of light shortage of a small number of sub-apertures. However, the method ignores the integrity of information, and the detection precision of the wavefront sensor is reduced along with the increase of the number of the apertures of the lacking photons. Therefore, it is necessary to find a method capable of estimating the centroid displacement of the missing photon aperture by using partial photon aperture data, so that the shack-hartmann wavefront sensor has a certain tolerance to the loss of molecular aperture information.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the problem that due to the fact that the sub-aperture of the shack-Hartmann wavefront sensor is lack of light, wavefront reconstruction accuracy is lowered is solved, the tolerance of the shack-Hartmann wavefront sensor to the sub-aperture lack of light is improved, and the influence of the sub-aperture lack of light on wavefront restoration accuracy and restoration algorithm robustness is reduced.
The technical scheme adopted by the invention for solving the technical problems is as follows: a photon-lacking aperture centroid displacement estimation method based on a neural network is characterized in that a nonlinear relation between local and full-aperture subaperture centroid displacements of a shack-Hartmann wavefront sensor is established based on the neural network, and the photon-lacking aperture centroid displacement is estimated according to the known partial subaperture centroid displacement, and is specifically realized through the following steps:
step 1: constructing a training set according to local subaperture centroid displacement and full-aperture subaperture centroid displacement of the shack-Hartmann wavefront sensor;
step 2: constructing a neural network which meets the corresponding relation between the local subaperture centroid displacement matrix and the full-aperture subaperture centroid displacement matrix;
and step 3: training the neural network constructed in the step 2 by using the training set generated in the step 1;
and 4, step 4: in practical application, the displacement of the mass center of the sub-aperture which can not be accurately detected is set to zero, namely a local displacement matrix of the mass center of the sub-aperture is obtained, then the displacement matrix is input into the neural network trained in the step 3 to obtain the displacement of the mass center of the full aperture sub-aperture, namely the displacement estimation of the mass center of the photon-lacking aperture is completed, and the sub-aperture which can not be accurately detected comprises the photon-lacking aperture and/or the sub-aperture with low signal to noise ratio.
Further, the training set constructing process described in step 1 is as follows:
step 1.1: the distorted wavefront is input into a shack-Hartmann wavefront sensing system, a light spot array image is output, and the full-aperture sub-aperture centroid is obtained through a centroid algorithmDisplacement matrix T x 、T y As the output of the neural network;
step 1.2: randomly selecting a plurality of sub-apertures as the default sub-apertures, and setting the sub-aperture centroid displacement of the corresponding position to be zero to obtain a local sub-aperture centroid displacement matrix L x 、L y The matrix being the input to the neural network, and T x 、T y Forming a group of training samples;
step 1.3: and (3) repeating the step 1.1 and the step 1.2, and randomly generating a plurality of groups of samples to form a training set.
Further, the neural network can be a BP neural network, a perceptron, a U-net network, a convolutional neural network, a deep neural network, or any other neural network satisfying input and output modes, and can also be other nonlinear fitting methods.
Further, the centroid algorithm may be a weighted centroid method (GoG), a weighted centroid method (WCoG), a windowing method, a threshold centroid method (TGoG), or any other centroid algorithm.
Further, in step 1.2, the input of the neural network is obtained by randomly assuming 10% to 50% of the effective sub-apertures as the missing photon apertures and zeroing the centroid shift of the sub-aperture at the corresponding position.
Compared with the prior art, the invention has the following advantages:
the nonlinear relation between the local part of the shack-Hartmann wavefront sensor and the full-aperture sub-aperture centroid displacement is established through the neural network, and the missing photon aperture centroid displacement is estimated through the partial sub-aperture centroid displacement information. Compared with the traditional photon-lacking aperture slope zero setting technology, the method effectively reduces the influence of the light lacking of the sub-aperture on the detection precision of the shack-Hartmann wavefront sensor and the robustness of the wavefront restoration algorithm, improves the tolerance of the wavefront sensor to the loss of sub-faculae, and can be used for high-precision wavefront restoration under the conditions of uneven light intensity distribution of the near-field of beacon light, flickering light intensity and the like.
Drawings
FIG. 1 is a diagram of a data acquisition device of the present invention.
Wherein 1: laser, 2, 3, 10, 11: relay lens, 4: aperture, 5: linear polarizer, 6: light-absorbing black block, 7: beam splitter prism, 8: spatial Light Modulator (SLM), 9: mirror, 12: microlens array, 13: CCD camera, 14: and (4) a computer.
Fig. 2 is a schematic diagram of a neural network model used in the embodiment.
FIG. 3 is a set of training samples of a neural network in an embodiment.
FIG. 4 is a statistical result of the prediction error of the neural network in the embodiment.
Wherein fig. 4(a) is root mean square statistics of the centroid displacement prediction error of each of the test samples, and fig. 4(b) is statistics of the maximum centroid displacement prediction error of each of the test samples, each of which contains 1000 sets of test samples under each of the light-lacking ratios.
FIG. 5 is a Root Mean Square (RMS) statistic of wavefront reconstruction residuals for the present invention and conventional methods in an example.
Wherein fig. 5(a) is the root mean square of the wavefront recovery residuals of all test sets, and fig. 5(b) is the statistical result of the root mean square of the wavefront recovery residuals of the test sets, each of which contains 1000 groups of test samples under each light-lacking ratio.
Detailed Description
In order that the objects and technical solutions of the present invention will be more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings in conjunction with specific embodiments.
In an embodiment, the apparatus system for collecting data is shown in fig. 1, wherein a continuous laser 1 is used as a light source, a relay lens 2 and a relay lens 3 constitute a relay system for collimating and expanding light beams, and further comprises an aperture 4, a linear polarizer 5, a light absorbing black block 6, a beam splitter prism 7, a 4-f system consisting of a pure phase Spatial Light Modulator (SLM)8, a mirror 9, a relay lens 10 and a relay lens 11, a micro lens array 12, a CCD camera 13, and a computer 14, wherein the surface of the spatial light modulator 8 and the front surface of the micro lens array 12 are made conjugate. In the embodiment, the data acquisition system is simulated, and the main parameters are set as follows: wavelength 635nm, number of sub-apertures 16 × 16, sub-aperture size 400 μm, microlens focal length 42.3mm, pixel size 6.4 μm × 6.4 μm. With the embodiment, the method specifically comprises the following steps:
Step 1: and constructing a training set according to the local subaperture centroid displacement and the full-aperture subaperture centroid displacement of the shack-Hartmann wavefront sensor.
The specific construction process of the training set in the step 1 is as follows:
step 1.1: generating a circular-caliber random incident wavefront containing the first 35 orders (except for translation and inclination terms) based on a Kolmogorov turbulence model, wherein the parameter D/r0 of the turbulence model is distributed between 5 and 10, D is the effective caliber of a shack-Hartmann wavefront sensor, and r0 is the atmospheric coherence length. The spatial light modulator is used for generating distorted wavefront, continuous light beams emitted by the laser are collimated and expanded by the relay lenses 2 and 3, then pass through the aperture 4 with the diameter of 6.4nm and the linear polaroid 5, generate polarized light required by the spatial light modulator, and vertically enter the spatial light modulator. After vertical reflection by the spatial light modulator, the light beam is converted into a phase modulation light beam, then the phase modulation light beam sequentially passes through a beam splitter prism 7, a reflector 9 and relay lenses 10 and 11 and enters a shack-Hartmann wavefront sensor (composed of a micro lens array 12 and a CCD 13), a light spot array image is formed on the CCD, the relay lenses 10 and 11 form a 4-f system, so that the surface of the spatial light modulator and the front surface of the micro lens array are conjugated, and finally, a computer obtains a full-aperture sub-aperture centroid displacement matrix T by a gravity center method x 、T y As an output of the network;
step 1.2: randomly assuming 10-50% of effective sub-apertures as the photon-lacking apertures, and setting the centroid displacement of the corresponding position to zero to obtain a local sub-aperture centroid displacement matrix L x 、L y As input to the network, with T x 、T y Forming a group of training samples;
step 1.3: and (3) repeating the step 1.1 and the step 1.2, and randomly generating 50000 groups of training samples to form a training set.
Step 2: and (3) constructing a neural network of the local subaperture centroid displacement and the full-aperture subaperture centroid displacement, and establishing a nonlinear relation between the local and full-aperture subaperture centroid displacements of the wavefront sensor. And estimating the missing photon aperture centroid displacement by the known sub-aperture centroid displacement based on the trained neural network. In the embodiment, an SH-Unet neural network model is constructed based on the U-net network, and the model structure is shown in FIG. 2. The network consists of an Encoder part and a Decoder part, wherein the Encoder part consists of two convolution layers and an active layer of 3 multiplied by 3, two max poling layers (stride 2) repeatedly, downsampling is carried out for 4 times to extract characteristic information, then upsampling is carried out for 4 times through the Decoder part symmetry, the Decoder part consists of an upsampling convolution layer and an active layer of 2 multiplied by 2, a Skip-Connection layer (the result of upsampling by feature map corresponding to the Encoder and the Decoder is added), two convolution layers and active layers of 3 multiplied by 3, and finally an image with the resolution same as the input size is output through fitting.
And 3, step 3: and (3) after the neural network is constructed, training the neural network by using the training set in the step (1), and storing the neural network after the training is finished.
And 4, step 4: inputting 1000 groups of distorted wavefronts into a shack-Hartmann wavefront sensing system to obtain a light spot array image, and solving the mass center displacement T of each sub-aperture by using a gravity center method testx And T testy Then respectively randomly selecting 10%, 20%, 30%, 40% and 50% of effective sub-apertures, and setting the centroid displacement to zero to obtain a local centroid displacement matrix L testx And L testy As test samples, 1000 test samples each having light shortage ratios of 10%, 20%, 30%, 40%, and 50% were used for 5000 test samples. Will L testx And L testy Inputting the neural network trained in the step 3 to obtain the mass center displacement of the full-aperture sub-aperture, namely completing the mass center displacement prediction of the light-lacking position, and combining the prediction result of the light-lacking position with T testx And T testy The real values in (1) are compared to calculate the prediction error, and the performance of the neural network is tested.
The statistical graph of the prediction results of the neural network on 5000 groups of test samples is shown in fig. 4, and each light-lacking ratio comprises 1000 groups of test samples, wherein fig. 4(a) is the root-mean-square statistics of the centroid displacement prediction error of the photon-lacking aperture of each test sample, and fig. 4(b) is the maximum centroid displacement prediction error statistics of the photon-lacking aperture of each sample. The centroid displacement Root Mean Square (RMS) error of the photon-lacking aperture is distributed in 0.0115-1.0868 pixels (pixels), and the average value is 0.0753 pixels. Meanwhile, in the embodiment, the input wavefront is restored based on an SH-Unet model and a traditional photon-lacking aperture slope zero setting method, and the restoration result is compared with the restoration result under the condition of no photon spot deletion. The wavefront residual error RMS result is shown in fig. 5, fig. 5(a) is the recovery result of 5000 groups of samples, the RMS distribution is between 0.0011 λ and 0.0559 λ, the mean value is 0.0074 λ, the difference is only 0.0015 λ from the ideal condition, and the recovery precision is improved by 93.32% compared with the traditional photon-lacking aperture slope zero setting method. Fig. 5(b) shows the statistical results of wavefront residuals RMS in different light-absence situations. As can be seen from the figure, the error of the traditional wave-front restoration algorithm increases with the increase of the light-lacking proportion, and the restoration error based on the SH-Unet model can be kept low under different light-lacking ratios. In conclusion, the method can accurately estimate the mass center displacement of the photon-lacking aperture according to the local sub-aperture mass center displacement data, improves the tolerance of the shack-Hartmann wavefront sensor to the loss of the sub-facula, and is expected to be used for high-precision wavefront measurement in the environments of uneven near-field light intensity, flickering light intensity and the like.
The above description is only an embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can understand that the modifications or substitutions within the technical scope of the present invention are included in the scope of the present invention.
Claims (4)
1. A missing photon aperture centroid displacement estimation method based on a neural network is characterized in that: the method is based on a neural network to establish a nonlinear relation between the local and full-aperture subaperture centroid displacements of a shack-Hartmann wavefront sensor, and estimates the lacking photon aperture centroid displacement according to the known partial subaperture centroid displacement, and is realized by the following steps:
step 1: constructing a training set according to local subaperture centroid displacement and full-aperture subaperture centroid displacement of the shack-Hartmann wavefront sensor;
in step 1, the training set construction process is as follows:
step 1.1: the distorted wavefront is input into a shack-Hartmann wavefront sensing system and outputGenerating a light spot array image, and obtaining a full-aperture sub-aperture centroid displacement matrix T through a centroid algorithm x 、T y As the output of the neural network;
step 1.2: randomly selecting a plurality of sub-apertures as the default sub-apertures, and setting the sub-aperture centroid displacement of the corresponding position to be zero to obtain a local sub-aperture centroid displacement matrix L x 、L y The matrix as input to the neural network, and T x 、T y Forming a group of training samples;
step 1.3: repeating the step 1.1 and the step 1.2, and randomly generating a plurality of groups of training samples to form a training set;
step 2: constructing a neural network which meets the corresponding relation between the local subaperture centroid displacement matrix and the full-aperture subaperture centroid displacement matrix;
and step 3: training the neural network constructed in the step 2 by using the training set generated in the step 1, and storing the neural network after the training is finished;
and 4, step 4: in practical application, the displacement of the mass center of the sub-aperture which can not be accurately detected is set to zero, namely a local displacement matrix of the mass center of the sub-aperture is obtained, then the displacement matrix is input into the neural network trained in the step 3 to obtain the displacement of the mass center of the full aperture sub-aperture, namely the displacement estimation of the mass center of the photon-lacking aperture is completed, and the sub-aperture which can not be accurately detected comprises the photon-lacking aperture and/or the sub-aperture with low signal to noise ratio.
2. The method for estimating the missing photon aperture centroid displacement based on the neural network as claimed in claim 1, wherein: the neural network may be a BP neural network, a perceptron, a U-net network, a convolutional neural network, a deep neural network, or any other neural network that satisfies the input and output patterns.
3. The method for estimating the missing photon aperture centroid displacement based on the neural network as claimed in claim 1, wherein: the centroid algorithm may be the weighted centroid method (GoG), or may be the weighted centroid method (WCoG), the windowed method, the threshold centroid method (TGoG), or any other centroid algorithm.
4. The method for estimating the missing photon aperture centroid displacement based on the neural network as claimed in claim 1, wherein: the input of the neural network in the step 1.2 is obtained by randomly assuming 10% -50% of effective sub-apertures as the missing photon apertures and zeroing the centroid displacement of the corresponding position.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110933080.6A CN113654670B (en) | 2021-08-12 | 2021-08-12 | Photon-lacking aperture centroid displacement estimation method based on neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110933080.6A CN113654670B (en) | 2021-08-12 | 2021-08-12 | Photon-lacking aperture centroid displacement estimation method based on neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113654670A CN113654670A (en) | 2021-11-16 |
| CN113654670B true CN113654670B (en) | 2022-07-29 |
Family
ID=78491599
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110933080.6A Active CN113654670B (en) | 2021-08-12 | 2021-08-12 | Photon-lacking aperture centroid displacement estimation method based on neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113654670B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114186664B (en) * | 2021-12-10 | 2023-09-19 | 中国科学院光电技术研究所 | Mode wavefront restoration method based on neural network |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110044498A (en) * | 2019-04-18 | 2019-07-23 | 中国科学院光电技术研究所 | A kind of Hartmann wave front sensor modal wavefront reconstruction method based on deep learning |
| CN110207835A (en) * | 2019-05-23 | 2019-09-06 | 中国科学院光电技术研究所 | A kind of wave front correction method based on out-of-focus image training |
| CN110646100A (en) * | 2019-09-30 | 2020-01-03 | 中国科学院光电技术研究所 | BP neural network-based frequency multiplication wavefront detection method |
| CN111579097A (en) * | 2020-05-21 | 2020-08-25 | 浙江大学 | High-precision optical scattering compensation method based on neural network |
| CN111626997A (en) * | 2020-05-21 | 2020-09-04 | 浙江大学 | Method for directly detecting optical distortion phase by high-speed single image based on deep learning |
-
2021
- 2021-08-12 CN CN202110933080.6A patent/CN113654670B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110044498A (en) * | 2019-04-18 | 2019-07-23 | 中国科学院光电技术研究所 | A kind of Hartmann wave front sensor modal wavefront reconstruction method based on deep learning |
| CN110207835A (en) * | 2019-05-23 | 2019-09-06 | 中国科学院光电技术研究所 | A kind of wave front correction method based on out-of-focus image training |
| CN110646100A (en) * | 2019-09-30 | 2020-01-03 | 中国科学院光电技术研究所 | BP neural network-based frequency multiplication wavefront detection method |
| CN111579097A (en) * | 2020-05-21 | 2020-08-25 | 浙江大学 | High-precision optical scattering compensation method based on neural network |
| CN111626997A (en) * | 2020-05-21 | 2020-09-04 | 浙江大学 | Method for directly detecting optical distortion phase by high-speed single image based on deep learning |
Non-Patent Citations (6)
| Title |
|---|
| Compressive Shack–Hartmann wavefront sensor based on deep neural networks;Peng Jia等;《MNRAS》;20210402;第503卷;第3194-3203页 * |
| Deep learning wavefront sensing method for Shack-Hartmann sensors with sparse sub-apertures;YULONG HE等;《Optics Express》;20210521;第29卷(第11期);第17669-17681页 * |
| High-precision wavefront reconstruction from Shack-Hartmann wavefront sensor data by a deep convolutional neural network;Hu Gu等;《Meas. Sci. Technol.》;20210511;第32卷;第085101-1-085101-8页 * |
| Performance of Centroiding Algorithms at Low Light Level Conditions in Adaptive Optics;Akondi Vyas等;《2009 International Conference on Advances in Recent Technologies in Communication and Computing》;20091117;第366-369页 * |
| 基于深度强化学习的自适应光学波前控制研究;许振兴;《中国博士学位论文全文数据库 基础科学辑》;20210315;正文第27页 * |
| 部分子孔径缺光对夏克－哈特曼波前传感器波前复原的影响;魏平等;《中国激光》;20200430;第47卷(第7期);第0409002-1-0409002-6页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113654670A (en) | 2021-11-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110044498B (en) | Hartmann wavefront sensor mode wavefront restoration method based on deep learning | |
| CN111221132B (en) | Method and device for measuring vortex beam topological charge number by fan-shaped sub-aperture micro-lens array | |
| CN111047681B (en) | Single-pixel three-dimensional end-to-end reconstruction method and device based on deep learning | |
| US8786864B2 (en) | Circular common-path point diffraction interference wavefront sensor | |
| CN104320598B (en) | A kind of aberration free image reconstructing method demarcated based on ccd array pixel response function frequency domain | |
| CN102564575B (en) | Laser far-field focal spot measuring method based on orthogonal optical wedge light splitting characteristics and focal spot reconstruction algorithm | |
| US7619191B1 (en) | Increase spatial sampling for wave front mid-spatial frequency error recovery | |
| CN112880986B (en) | Spliced telescope translation error detection method based on convolutional neural network | |
| CN106644105A (en) | Wavefront sensor based on double helix point spread function, detection method, and system | |
| CN113654670B (en) | Photon-lacking aperture centroid displacement estimation method based on neural network | |
| CN114186664B (en) | Mode wavefront restoration method based on neural network | |
| CN111579097A (en) | High-precision optical scattering compensation method based on neural network | |
| CN114323310B (en) | High-resolution Hartmann wavefront sensor | |
| CN116429252A (en) | Comprehensive parameter evaluation method for laser beam quality | |
| CN111998962B (en) | Hartmann wavefront sensor based on array type binary phase modulation | |
| CN103105235B (en) | Method for measuring oblique aberration in sub-aperture of Hartmann wavefront sensor | |
| CN210893429U (en) | Defocusing type light field camera wavefront sensor | |
| CN112097681A (en) | Complex optical curved surface shape error detection method based on speckle field phase recovery | |
| CN100378442C (en) | Optical wave front detection device and detection method thereof | |
| CN115294422A (en) | Extended target super-resolution wavefront restoration method | |
| CN117451189A (en) | Wavefront detection method based on Hartmann detector | |
| CN116124304A (en) | Neural network mode wavefront restoration method based on limited sampling information | |
| CN118670691B (en) | Imaging quality measuring device and measuring method for large-caliber optical imaging system | |
| Soria et al. | Study of Three WFS for the Modular System in a Portable AO Instrument: ALIOLI. | |
| CN118225256A (en) | Hartmann detector mode restoration method based on least square correction restoration matrix |
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 |