CN114660917B - Free-field infrared digital holographic imaging method based on Transformer model - Google Patents
Free-field infrared digital holographic imaging method based on Transformer model Download PDFInfo
- Publication number
- CN114660917B CN114660917B CN202210276569.5A CN202210276569A CN114660917B CN 114660917 B CN114660917 B CN 114660917B CN 202210276569 A CN202210276569 A CN 202210276569A CN 114660917 B CN114660917 B CN 114660917B
- Authority
- CN
- China
- Prior art keywords
- field
- sub
- sample
- translation table
- images
- 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
- 238000003384 imaging method Methods 0.000 title claims description 20
- 238000000034 method Methods 0.000 claims abstract description 32
- 238000012549 training Methods 0.000 claims abstract description 18
- 238000010276 construction Methods 0.000 claims abstract 2
- 238000013519 translation Methods 0.000 claims description 60
- 238000004088 simulation Methods 0.000 claims description 13
- BQCADISMDOOEFD-UHFFFAOYSA-N Silver Chemical compound [Ag] BQCADISMDOOEFD-UHFFFAOYSA-N 0.000 claims description 12
- 229910052709 silver Inorganic materials 0.000 claims description 12
- 239000004332 silver Substances 0.000 claims description 12
- 230000001681 protective effect Effects 0.000 claims description 11
- 238000004422 calculation algorithm Methods 0.000 claims description 9
- 238000010586 diagram Methods 0.000 claims description 9
- 239000011159 matrix material Substances 0.000 claims description 9
- 238000009499 grossing Methods 0.000 claims description 7
- 238000010587 phase diagram Methods 0.000 claims description 7
- 238000006073 displacement reaction Methods 0.000 claims description 6
- 238000012856 packing Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000008878 coupling Effects 0.000 claims description 3
- 238000010168 coupling process Methods 0.000 claims description 3
- 238000005859 coupling reaction Methods 0.000 claims description 3
- 230000008030 elimination Effects 0.000 claims description 3
- 238000003379 elimination reaction Methods 0.000 claims description 3
- 238000007526 fusion splicing Methods 0.000 claims description 3
- 239000011521 glass Substances 0.000 claims description 3
- 238000003709 image segmentation Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 239000013307 optical fiber Substances 0.000 claims description 3
- 238000003672 processing method Methods 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- 238000005094 computer simulation Methods 0.000 claims 1
- 230000000007 visual effect Effects 0.000 description 3
- 238000012937 correction Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000012528 membrane Substances 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 230000002194 synthesizing effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G03—PHOTOGRAPHY; CINEMATOGRAPHY; ANALOGOUS TECHNIQUES USING WAVES OTHER THAN OPTICAL WAVES; ELECTROGRAPHY; HOLOGRAPHY
- G03H—HOLOGRAPHIC PROCESSES OR APPARATUS
- G03H1/00—Holographic processes or apparatus using light, infrared or ultraviolet waves for obtaining holograms or for obtaining an image from them; Details peculiar thereto
- G03H1/04—Processes or apparatus for producing holograms
- G03H1/0443—Digital holography, i.e. recording holograms with digital recording means
-
- G—PHYSICS
- G03—PHOTOGRAPHY; CINEMATOGRAPHY; ANALOGOUS TECHNIQUES USING WAVES OTHER THAN OPTICAL WAVES; ELECTROGRAPHY; HOLOGRAPHY
- G03H—HOLOGRAPHIC PROCESSES OR APPARATUS
- G03H1/00—Holographic processes or apparatus using light, infrared or ultraviolet waves for obtaining holograms or for obtaining an image from them; Details peculiar thereto
- G03H1/04—Processes or apparatus for producing holograms
- G03H1/08—Synthesising holograms, i.e. holograms synthesized from objects or objects from holograms
- G03H1/0866—Digital holographic imaging, i.e. synthesizing holobjects from holograms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- Holo Graphy (AREA)
Abstract
The invention discloses a step one of constructing a real sample data set collection system; step two, collecting a real sample data set of the Transformer model; step three, training set generation and Transformer model construction; fourthly, eliminating background noise based on a Transformer model and fusing and splicing images; and step five, dynamically recording the long time axis of the sample. The invention has the advantages that: the method has the advantages that the focusing distance judgment accuracy can be effectively improved, the obvious crease existing in the final synthesized main view field amplitude image and main view field phase image is eliminated, and the high-resolution main view field amplitude image and the main view field phase image without the obvious crease in a period of continuous time are shot.
Description
Technical Field
The invention relates to the technical field of infrared digital holographic imaging, in particular to a free-field infrared digital holographic imaging method based on Transformer model error correction.
Background
The infrared band is outside the visible band, so that the infrared band has good application prospect in the imaging field due to different penetrability to different substances, and the digital holographic technology has unique advantage in imaging a plurality of focusing layers of a specific target sample due to a simple imaging structure and a flexible data processing mode. The Transformer model Is the seq2seq model proposed in the paper Attention Is All You Need published by google at the end of 2017. The feature of the transducer seq2seq model is that a lot of attention layers are used in the model. The Transformer model now also shows very satisfactory results in computer vision. In the traditional imaging process, the problem that the focusing distance judgment accuracy is low and obvious creases exist in a final synthesized main view field amplitude image and a final synthesized main view field phase image exists, the training and the recognition of infrared digital holographic imaging are realized through a Transformer model, and the free view field infrared digital holographic imaging technology based on the Transformer model correction can be realized by combining the advantages of the Transformer model and the main view field phase image.
The existing free visual field infrared digital holographic imaging method has the following defects: the accuracy of judging the focusing distance is low, and obvious creases exist in the final synthesized main view field amplitude image and the main view field phase image, so that the image quality is seriously influenced.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a free-field infrared digital holographic imaging method based on a Transformer model. The problem of focus distance judge the precision low and the final main visual field amplitude image of synthesizing and main visual field phase place image have obvious crease is solved.
In order to realize the purpose of the invention, the technical scheme adopted by the invention is as follows:
a free-field infrared digital holographic imaging method based on a Transformer model comprises the following steps:
step one, constructing a real sample data set collection system;
the collection system includes: tunable laser light source 1, protection silver membrane reflective collimator 2, two-dimensional electronic translation platform 3, sample rotary translation platform 4, infrared pyroelectric detector 5 and detector two-dimensional translation platform 6.
The tunable laser light source 1 is used for emitting infrared laser and is connected to the protective silver film reflective collimator 2 through an optical fiber; the protective silver film reflective collimator 2 is used for coupling infrared laser with maximum efficiency through a sample, and the two-dimensional electric translation stage 3 and the sample rotation translation stage 4 are matched with each other to enable the sample to translate and rotate; the two-dimensional electric translation stage 3 and the sample rotary translation stage 4 are positioned right below the protective silver film reflective collimator 2. The infrared pyroelectric detector 5 is used for receiving infrared laser thermal signals and is positioned under the two-dimensional electric translation table 3. The detector two-dimensional translation table 6 is positioned below the infrared pyroelectric detector 5 and can enable the infrared pyroelectric detector 5 to translate.
And step two, a real sample data set collection process of the transform model is carried out, sample data acquisition is completed through a free-field-of-view infrared digital holographic imaging system, the slide glass coated with the sample is placed on a two-dimensional electric translation table 3 on a sample rotation translation table 4, the horizontal scanning times C, the longitudinal scanning times Z, the scanning path and the stepping distance of the two-dimensional electric translation table 3 are set, and the rotation times X, the clockwise direction and the stepping degrees 360/(X + 1) DEG of the sample rotation translation table 4 are set. The sample free field hologram collection process adopts a traversing mode, and the infrared pyroelectric detector 5 collects C +1 sub-field holograms through C times of horizontal movement. And then the two-dimensional electric translation stage 3 completes longitudinal movement Z times, translates to the initial row position every time of longitudinal movement, and repeats horizontal movement scanning C times. And then, the sample rotation translation stage 4 finishes X times of rotation scanning, and horizontal movement scanning and longitudinal movement scanning are carried out on each rotation displacement, so that (X + 1) × (Z + 1) × (C + 1) sub-field holograms are collected.
And thirdly, generating a training set and constructing a Transformer model, simulating a hologram record and a reconstruction process through a computer to obtain a corresponding training set, simulating according to a simulation sample to obtain a simulation sub-field-of-view hologram, and preprocessing, wherein the preprocessing process comprises the operations of adding different levels of noise, rotating an image and stretching the image to the simulation sub-field-of-view hologram. Converting the preprocessed image of the sub-field hologram into a data set form to obtain a training set which is composed of simulation data and used for reconstructing the sub-field hologram, setting a self-attention layer and a multi-head attention layer in a Transformer model and packing the self-attention layer and the multi-head attention layer into a matrix Q, packing keys and values into matrixes K and V, wherein a matrix calculation formula is as follows:
MultiHead(Q,K,V)=Concat(head 1 ,...,head n )W O
head i =Attention(QW i Q ,KW i K ,VW i V )
the softmax function is to take values in the matrix according to a certain weight, the Concat function is to connect all the internal matrices, train the preprocessed data set to obtain a classified result, and realize the focusing distance judgment of the reconstructed amplitude diagram and phase diagram of the sub-field. And finally, (X + 1) ((Z + 1) ((C + 1)) pieces of sub-field holograms in the step one are used as input layers, and (X + 1) ((Z + 1) ((C + 1)) pieces of optimized sub-field amplitude images and sub-field phase images are obtained on the output layers.
And fourthly, performing background noise elimination and image fusion splicing process based on a Transformer model, training a smoothing algorithm based on the principle of the Transformer model, eliminating small errors when the two-dimensional electric translation stage 3 and the sample rotary translation stage 4 are displaced by the smoothing algorithm, and calibrating the background noise of the sub-field-of-view holograms at different positions to the same level. And the smoothing algorithm is used for carrying out normalization processing on the subfield images. And (2) carrying out image segmentation by using the (X + 1) ((Z + 1)) (C + 1) processed sub-field amplitude images and sub-field phase images in the second step, adding random noise interference in the segmentation process, obtaining a training set according to the data processing method in the second step, setting a self-attention layer and a multi-head attention layer of a transform model according to the method in the second step, realizing splicing and restoring of the sub-field images in the training set, obtaining an optimal background noise value, eliminating displacement errors generated by the two-dimensional sample electric translation stage and the sample rotary translation stage, and obtaining a normalized result. And finally, taking the (X + 1) (Z + 1) ((C + 1)) processed sub-field amplitude images and sub-field phase images in the second step as input layers, and obtaining the intensity-calibrated and smoothed main field amplitude images and main field phase images on the output layers.
And step five, dynamically recording the sample long time axis, realizing the whole process, recording and time information calibration of the sample image through the step two, and realizing the complete display of the object change holographic information in a long time through the stacking of the main view field holograms obtained in the step four according to the time sequence.
Preferably, in the process of stacking in time sequence in the fifth step, the time gap may be set to take one sub-field amplitude map and one sub-field phase map every N1, and record N2 hours, so as to realize N3 main field amplitude maps and main field phase maps in total in the long time axis within N2 hours. Wherein N1, N2 and N3 are all integer parameters which can be set.
Compared with the prior art, the invention has the advantages that:
the method can effectively improve the judgment accuracy of the focusing distance, eliminate obvious creases in the final synthesized main view field amplitude image and main view field phase image, and shoot the main view field amplitude image and the main view field phase image which have high resolution and do not have obvious creases in a period of continuous time.
Drawings
FIG. 1 is a flow chart of a free-field infrared digital holographic imaging method according to an embodiment of the invention;
FIG. 2 is a schematic diagram of a real sample data set collection system according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of an image collected by a real sample data set collection system according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be further described in detail below with reference to the accompanying drawings by way of examples.
As shown in fig. 1, a method for free-field infrared digital holographic imaging based on a Transformer model includes the following steps:
step one, as shown in fig. 2, a real sample dataset collection system is constructed, the collection system comprising: the device comprises a 1550nm tunable laser light source 1, a protective silver film reflective collimator 2, a two-dimensional electric translation table 3, a sample rotation translation table 4, an infrared pyroelectric detector 5 and a detector two-dimensional translation table 6.
The tunable laser light source 1 is used for emitting infrared laser and is connected to the protective silver film reflective collimator 2 through an optical fiber; the protective silver film reflective collimator 2 is used for coupling infrared laser with maximum efficiency through a sample, and the two-dimensional electric translation stage 3 and the sample rotation translation stage 4 are matched with each other to enable the sample to translate and rotate; the two-dimensional electric translation stage 3 and the sample rotation translation stage 4 are positioned right below the protective silver film reflective collimator 2. The infrared pyroelectric detector 5 is used for receiving infrared laser thermal signals and is positioned under the two-dimensional electric translation table 3. The detector two-dimensional translation table 6 is positioned below the infrared pyroelectric detector 5 and can enable the infrared pyroelectric detector 5 to translate.
And step two, a real sample data set collection process of the transform model is carried out, sample data acquisition is completed through a free-field infrared digital holographic imaging system, the glass slide coated with the sample is placed on a two-dimensional electric translation table 3 of a sample rotation translation table 4, the two-dimensional electric translation table 3 is set to have two horizontal scanning times, two longitudinal scanning times, a scanning path and a stepping distance, and the sample rotation translation table 4 is set to have three rotation times, a clockwise rotation direction and 90 degrees of stepping degrees. The sample free field hologram acquisition process adopts a traversing mode, and the infrared pyroelectric detector 5 acquires three sub-field holograms through two times of horizontal movement. And then the two-dimensional electric translation stage 3 completes longitudinal movement twice, translates to the initial position every time of longitudinal movement, and repeats horizontal movement scanning twice. And then, completing rotational scanning 3 times by the sample rotational translation stage 4, and performing horizontal movement scanning and longitudinal movement scanning on each rotational displacement to acquire 4 × 3 sub-field holograms.
And thirdly, generating a training set and constructing a Transformer model, simulating a hologram record and a reconstruction process through a computer to obtain a corresponding training set, simulating according to a simulation sample to obtain a simulation sub-field-of-view hologram, and preprocessing, wherein the preprocessing process comprises the operations of adding different levels of noise, rotating an image and stretching the image to the simulation sub-field-of-view hologram. Converting the preprocessed image of the sub-field hologram into a data set form to obtain a training set which is composed of simulation data and used for reconstructing the sub-field hologram, setting a self-attention layer and a multi-head attention layer in a Transformer model and packing the self-attention layer and the multi-head attention layer into a matrix Q, packing keys and values into matrixes K and V, wherein a matrix calculation formula is as follows:
MultiHead(Q,K,V)=Concat(head 1 ,...,head n )W O
head i =Attention(QW i Q ,KW i K ,VW i V )
the softmax function is to take values in the matrix according to a certain weight, the Concat function is to connect all the internal matrices, train the preprocessed data set to obtain a classified result, and realize the focusing distance judgment of the reconstructed amplitude diagram and phase diagram of the sub-field. And finally, taking 4 × 3 subfields holograms in the step one as an input layer, and obtaining 4 × 3 optimized subfields amplitude images and subfields phase images on an output layer.
And fourthly, the background noise elimination and image fusion splicing process based on the transform model is limited by the low sensitivity of the infrared detector and the movement error of the translation stage, and the background noise of the images collected at different positions has certain difference, so that the image quality is very low, and the target to be detected cannot be clearly identified. After the sub-field holograms are synthesized into the main field hologram, obvious creases exist at the junctions of the sub-field holograms in the image. The smoothing algorithm is trained through a principle based on a Transformer model, and because the two-dimensional electric translation stage 3 and the sample rotary translation stage 4 can not necessarily rotate to accurate positions when moving, the two-dimensional electric translation stage 3 and the sample rotary translation stage 4 can generate small errors, the algorithm can eliminate the small errors, the background noise of the sub-field-of-view holograms at different positions is calibrated to the same level, and the image quality is enhanced while the image splitting sense is eliminated. Because the intensity of the sub-field hologram is different due to the fact that the sample is located at different positions, the algorithm can perform normalization processing on the sub-field image. And (3) carrying out image segmentation by using the 4 × 3 processed sub-field amplitude images and the sub-field phase images in the second step, adding random noise interference in the segmentation process, obtaining a training set according to the data processing method in the second step, setting a self-attention layer and a multi-head attention layer of a transform model according to the method in the second step, realizing splicing and restoring of the sub-field images in the training set, solving an optimal background noise value, eliminating displacement errors generated by the two-dimensional sample electric translation stage and the two-dimensional sample rotary translation stage, and obtaining a normalized result. And finally, taking the 4 × 3 processed sub-field amplitude images and the sub-field phase images in the second step as input layers, and obtaining the main field amplitude images and the main field phase images which are smooth after intensity calibration on the output layers.
And step five, dynamically recording the long time axis of the sample, realizing the whole process, recording and time information calibration of the sample image through the step two, and realizing the complete display of the object change holographic information in a long time through the stacking of the main view field holograms obtained in the step four according to the time sequence, wherein in the process of stacking according to the time sequence, the time interval can be set to shoot a sub view field amplitude diagram and a sub view field phase diagram every 5s, and recording for 5 hours, so that the total 100 main view field amplitude diagrams and main view field phase diagrams in the 5-hour long time axis are realized.
It will be appreciated by those of ordinary skill in the art that the examples described herein are intended to assist the reader in understanding the practice of the invention, and it is to be understood that the scope of the invention is not limited to such specific statements and examples. Those skilled in the art can make various other specific changes and combinations based on the teachings of the present invention without departing from the spirit of the invention, and these changes and combinations are within the scope of the invention.
Claims (2)
1. A free field of view infrared digital holographic imaging method based on a Transformer model is characterized by comprising the following steps:
step one, constructing a real sample data set collection system;
the collection system includes: the device comprises a tunable laser light source (1), a protective silver film reflective collimator (2), a two-dimensional electric translation table (3), a sample rotation translation table (4), an infrared pyroelectric detector (5) and a detector two-dimensional translation table (6);
the tunable laser light source (1) is used for emitting infrared laser and is connected to the protective silver film reflective collimator (2) through an optical fiber; the protective silver film reflective collimator (2) is used for coupling infrared laser with maximum efficiency through a sample, and the two-dimensional electric translation table (3) and the sample rotation translation table (4) are matched with each other to enable the sample to translate and rotate; the two-dimensional electric translation stage (3) and the sample rotation translation stage (4) are positioned right below the protective silver film reflective collimator (2); the infrared pyroelectric detector (5) is used for receiving infrared laser thermal signals and is positioned right below the two-dimensional electric translation table (3); the detector two-dimensional translation table (6) is positioned below the infrared pyroelectric detector (5) and can enable the infrared pyroelectric detector (5) to translate;
step two, a real sample data set collection process of a transform model is carried out, sample data acquisition is completed through a free-field infrared digital holographic imaging system, a slide glass coated with a sample is placed on a two-dimensional electric translation table (3) on a sample rotation translation table (4), the horizontal scanning times C, the longitudinal scanning times Z, a scanning path and a stepping distance of the two-dimensional electric translation table (3) are set, the rotation times X, the clockwise direction and the stepping degrees 360/(X + 1) °of the sample rotation translation table (4) are set, a traversing mode is adopted in the sample free-field hologram collection process, C +1 sub-field holograms are collected through C horizontal movement, and an infrared pyroelectric detector (5) collects C +1 sub-field holograms; then the two-dimensional electric translation stage (3) completes longitudinal movement for Z times, each time the longitudinal movement is translated to the initial position, C times of horizontal movement scanning are repeated, then the sample rotation translation stage (4) completes rotation scanning for X times, each time the rotation displacement is performed with horizontal movement scanning and longitudinal movement scanning, and (X + 1) (Z + 1) ((C + 1)) sub-field holograms are collected together;
step three, training set generation and a Transformer model construction process, a corresponding training set is obtained through computer simulation hologram recording and reconstruction processes, a simulation sub-field-of-view hologram is obtained according to simulation sample simulation and is preprocessed, and the preprocessing process comprises operations of adding different levels of noise, image rotation and image stretching to the simulation sub-field-of-view hologram; converting the preprocessed image of the sub-field hologram into a data set form to obtain a training set which is composed of simulation data and used for reconstructing the sub-field hologram, setting a self-attention layer and a multi-head attention layer in a Transformer model and packing the self-attention layer and the multi-head attention layer into a matrix Q, packing keys and values into matrixes K and V, wherein a matrix calculation formula is as follows:
MultiHead(Q,K,V)=Concat(head 1 ,...,head n )W O
head i =Attention(QW i Q ,KW i K ,VW i V )
the softmax function is to take values in the matrix according to a certain weight, the Concat function is to connect all the internal matrices, train the preprocessed data set to obtain a classified result, and judge the focusing distance of the reconstructed amplitude diagram and phase diagram of the sub-field; finally, (X + 1) ((Z + 1) ((C + 1)) pieces of sub-field holograms in the step one are used as input layers, and (X + 1) ((Z + 1) ((C + 1)) pieces of optimized sub-field amplitude images and sub-field phase images are obtained on output layers;
fourthly, a background noise elimination and image fusion splicing process based on a Transformer model is carried out, a smoothing algorithm is trained through a principle based on the Transformer model, tiny errors generated when the two-dimensional electric translation table (3) and the sample rotation translation table (4) are displaced are eliminated through the smoothing algorithm, and background noises of the sub-field holograms at different positions are calibrated to be at the same level; the smoothing algorithm is used for carrying out normalization processing on the subfield images; performing image segmentation by using (X + 1) (Z + 1) (+ C + 1) processed sub-field amplitude images and sub-field phase images in the second step, adding random noise interference in the segmentation process, obtaining a training set according to the data processing method in the second step, setting a self-attention layer and a multi-head attention layer of a transform model according to the method in the second step, realizing splicing and restoring of the sub-field images in the training set, solving an optimal background noise value, eliminating displacement errors generated by a two-dimensional sample electric translation stage and a sample rotary translation stage, and obtaining a normalized result; finally, (X + 1) ((Z + 1) × (C + 1)) processed sub-field amplitude images and sub-field phase images in the second step are used as input layers, and the main field amplitude images and the main field phase images which are smoothed after intensity calibration are obtained on the output layers;
and step five, dynamically recording the sample long time axis, realizing the whole process, recording and time information calibration of the sample image through the step two, and realizing the complete display of the object change holographic information in a long time through the stacking of the main view field holograms obtained in the step four according to the time sequence.
2. The transform model-based free-field infrared digital holographic imaging method according to claim 1, wherein: in the process of stacking according to the time sequence in the fifth step, the time gap can be set to be one sub-field amplitude diagram and one sub-field phase diagram taken every N1, and N2 hours are recorded, so that N3 main field amplitude diagrams and main field phase diagrams are obtained in total within N2 hours on a long time axis, wherein N1, N2 and N3 are all settable integer parameters.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210276569.5A CN114660917B (en) | 2022-03-21 | 2022-03-21 | Free-field infrared digital holographic imaging method based on Transformer model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210276569.5A CN114660917B (en) | 2022-03-21 | 2022-03-21 | Free-field infrared digital holographic imaging method based on Transformer model |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114660917A CN114660917A (en) | 2022-06-24 |
CN114660917B true CN114660917B (en) | 2022-10-21 |
Family
ID=82030731
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210276569.5A Active CN114660917B (en) | 2022-03-21 | 2022-03-21 | Free-field infrared digital holographic imaging method based on Transformer model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114660917B (en) |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103838124A (en) * | 2014-01-21 | 2014-06-04 | 中国科学院大学 | Imaging-view-field-increase-oriented lamination scanning digital holography |
CN107145052A (en) * | 2017-05-24 | 2017-09-08 | 上海交通大学 | Holographic microscopic imaging method based on digital interpolative and phase iteration |
CN109188881A (en) * | 2018-10-12 | 2019-01-11 | 中国地质大学(北京) | A kind of THz wave digital hologram imaging method and system |
CN110297418A (en) * | 2019-07-11 | 2019-10-01 | 中国地质大学(北京) | A kind of THz wave digital hologram imaging method decomposed based on Terahertz diffraction pattern |
WO2021059909A1 (en) * | 2019-09-27 | 2021-04-01 | オムロン株式会社 | Data generation system, learning device, data generation device, data generation method, and data generation program |
CN112666814A (en) * | 2020-12-26 | 2021-04-16 | 北京工业大学 | Off-axis digital holographic diffraction tomography method based on continuous terahertz waves |
CN113093499A (en) * | 2021-04-15 | 2021-07-09 | 中国地质大学(北京) | Discrete aperture interpolation terahertz digital holographic imaging method and system |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2013517829A (en) * | 2010-01-22 | 2013-05-20 | コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ | Magnetic susceptibility gradient mapping |
EP2515136A1 (en) * | 2011-04-21 | 2012-10-24 | Koninklijke Philips Electronics N.V. | Contrast enhanced magnetic resonance angiography with chemical shift encoding for fat suppression |
-
2022
- 2022-03-21 CN CN202210276569.5A patent/CN114660917B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103838124A (en) * | 2014-01-21 | 2014-06-04 | 中国科学院大学 | Imaging-view-field-increase-oriented lamination scanning digital holography |
CN107145052A (en) * | 2017-05-24 | 2017-09-08 | 上海交通大学 | Holographic microscopic imaging method based on digital interpolative and phase iteration |
CN109188881A (en) * | 2018-10-12 | 2019-01-11 | 中国地质大学(北京) | A kind of THz wave digital hologram imaging method and system |
CN110297418A (en) * | 2019-07-11 | 2019-10-01 | 中国地质大学(北京) | A kind of THz wave digital hologram imaging method decomposed based on Terahertz diffraction pattern |
WO2021059909A1 (en) * | 2019-09-27 | 2021-04-01 | オムロン株式会社 | Data generation system, learning device, data generation device, data generation method, and data generation program |
CN112666814A (en) * | 2020-12-26 | 2021-04-16 | 北京工业大学 | Off-axis digital holographic diffraction tomography method based on continuous terahertz waves |
CN113093499A (en) * | 2021-04-15 | 2021-07-09 | 中国地质大学(北京) | Discrete aperture interpolation terahertz digital holographic imaging method and system |
Non-Patent Citations (1)
Title |
---|
基于相位拼接的数字全息大视场成像方法研究;张鑫等;《现代电子技术》;20131101;第第36卷卷(第21期);96-99 * |
Also Published As
Publication number | Publication date |
---|---|
CN114660917A (en) | 2022-06-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103201648B (en) | Calibration of a probe in ptychography | |
CN1985188B (en) | High resolution imaging | |
CN102239426B (en) | Provision of image data | |
CN111047681B (en) | Single-pixel three-dimensional end-to-end reconstruction method and device based on deep learning | |
CN107655405B (en) | Method for eliminating axial distance error between object and CCD by using self-focusing iterative algorithm | |
CN111366557A (en) | Phase imaging method based on thin scattering medium | |
CN110490818B (en) | Computed ghost imaging reconstruction recovery method based on CGAN | |
CN109900355B (en) | Imaging method and device | |
CN111650738A (en) | Fourier laminated microscopic image reconstruction method and device based on deep learning | |
CN110455834A (en) | X-ray single exposure imaging device and method based on light intensity transmission equation | |
EP4260026A1 (en) | Totagraphy: coherent diffractive/digital information reconstruction by iterative phase recovery using special masks | |
CN109597291A (en) | A kind of optical scanner hologram image recognition methods based on convolutional neural networks | |
CN114660917B (en) | Free-field infrared digital holographic imaging method based on Transformer model | |
CN114627017A (en) | Point cloud denoising method based on multi-level attention perception | |
CN111259991B (en) | Under-sampling single-pixel imaging target identification method in noise environment | |
CN114241072A (en) | Laminated imaging reconstruction method and system | |
CN115631253A (en) | Optical fiber lens-free generalized imaging method based on physical drive deep learning | |
WO2023023961A1 (en) | Piv image calibration apparatus and method based on laser linear array | |
CN115616014B (en) | Momentum-encoding X-ray diffraction pattern matching correction method and application thereof | |
CN108007387B (en) | Surface shape measurement device and method based on Structured Illumination | |
CN110160663A (en) | A kind of high-resolution near field Wavefront measuring apparatus and measurement method | |
CN115494005A (en) | Semiconductor defect detection device and method based on infrared microscopic digital holography | |
CN112348947B (en) | Three-dimensional reconstruction method for deep learning based on reference information assistance | |
CN112687009B (en) | Three-dimensional face representation method and parameter measurement device and method thereof | |
CN114998469A (en) | Regional infrared digital holographic method based on neural convolutional network |
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 |