CN104410789A - Staring super-resolution imaging device and method - Google Patents
Staring super-resolution imaging device and method Download PDFInfo
- Publication number
- CN104410789A CN104410789A CN201410746361.0A CN201410746361A CN104410789A CN 104410789 A CN104410789 A CN 104410789A CN 201410746361 A CN201410746361 A CN 201410746361A CN 104410789 A CN104410789 A CN 104410789A
- Authority
- CN
- China
- Prior art keywords
- detector
- image
- resolution
- super
- imaging device
- 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
Landscapes
- Transforming Light Signals Into Electric Signals (AREA)
Abstract
The invention provides a staring super-resolution imaging device and method, and aims at solving the problem of low resolution of an existing imaging device. The device comprises an imaging device lens group (1), a detector (2), a detector driving platform (3), a driver (4), a storage device (5), an image processing circuit (6) and an image display circuit (7), wherein the detector (2) is arranged in a focal plane on a light path of the imaging device lens group (1); a drive part of the driver (4) drives the detector driving platform (3) to generate movement with magnitude and direction being random values. The imaging method comprises the following realization steps: a change number instruction is set; the driver (4) drives the detector (2) to do corresponding number of random jitter; a light signal is imaged on the detector (2) through the imaging device lens group (1) to obtain a low-resolution image with a frame number corresponding to a change number; the low-resolution image is rebuilt by the utilization of a variational Bayesian algorithm. The staring super-resolution imaging device and method are suitable for video reconstruction and satellite photography.
Description
Technical field
The invention belongs to super-resolution imaging technical field, be specifically related to gazing type super-resolution imaging device and the formation method of the acquisition of a kind of multiple image and reconstruction, can be used for video reconstruction and satellite shooting.
Background technology
The resolution improving optical imaging system is the unremitting pursue of carrying out image science research and engineer applied thereof, except passing through to improve imaging system associated components, as optical system focal length, the performance of aperture and detector etc. is to improve outside resolution, can also select to add suitable optics in imaging systems to improve resolution, but be adjusted to the significantly increase that can cause the complicated of system and task as system configuration to a certain extent, thus how recover based on existing imaging device or reconstruct the active demand that super-resolution image becomes current numerous image applications field.In scientific research, people sight is put into gradually application contemporary optics and digital image processing techniques in the resolution improving imaging system.
Improve imaging device resolution to be realized by the reconstruct of single-frame images or multiple image, the amount of information comprised due to single-frame images is limited, lacks new information in process of reconstruction, and the effect that thus resolution improves is not very desirable.Based on the super-resolution rebuilding algorithm of multiple image, namely utilize the multiframe low-resolution image of Same Scene to obtain a frame super-resolution image of this scene, because the amount of information that multiple image comprises is greater than single-frame images, including similar but incomplete same complementary information and certain prior information between image sequence each other, thus providing possibility for recovering authentic and valid super-resolution image.
At present on the obtain manner of multiple image, three kinds of modes can be adopted to realize: the first obtains multiframe low-resolution image by the mobile of camera lens.Shenzhen Yatu Digital Video Frequency Technology Co., Ltd is in the application documents of 201010505956.9 at number of patent application, disclose the system that a kind of scioptics shifting method obtains multiframe low-resolution image, this Systematical control is simple, displacement is accurately adjustable, but drive circuit is more complicated, optical design is subject to micro-displacement mechanism restriction, and versatility is poor; The second adds parallel flat in light path, the acquisition of multiple image is realized by swing or Rotating Plates position, BJ University of Aeronautics & Astronautics is in the patent application of application number 201210451785.5, disclose a kind of parallel flat that adds in the optical path to obtain the method for multiple image, the method can realize the effect of dynamic scene super-resolution real time imagery, and this is practical mode.But it is very high to the requirement on machining accuracy of parallel flat, the simultaneously change of light path can cause the significantly increasing of the complicated of system and task amount, thus high cost in actual applications, and difficulty is excessive, and pre-set the frame number collecting Same Scene, limit the raising of resolution; The third obtains multiple image by the mode of mobile detector, Tsing-Hua University is in the patent application of 200810056002.7 at application number, disclose a kind of device by adopting the method for mobile detector to obtain multiple image, by controlling image rotation mechanism or making the micro-rotation of photodetector array obtain a series ofly having relative to micro-low-resolution image rotated.This device is compared with first kind of way, and practicality is high; Compared with the second way, because it does not introduce new optical element can not cause the complicated of system, manufacturing cost is lower.But precisely controlling displacement quantity in the third mode, high to hardware requirement, and in advance the micro-anglec of rotation of setting defines the scope of the multiple image of acquisition and collects the frame number of image of Same Scene, can cause that to rebuild effect undesirable.
Summary of the invention
The object of the invention is to overcome the defect of poor universality that prior art exists, complex process, reconstruction weak effect, provide a kind of gazing type super-resolution imaging device and method, by random mobile detector to obtain multiple image, for solving the low problem of existing imaging device resolution.
To achieve these goals, technical scheme of the present invention comprises:
Imaging device lens group 1, for gathering light signal, obtains the light signal collected;
Detector 2, for receiving the light signal collected, imaging thereon, and become image is exported to memory cell 5, image processing circuit 6 and image displaying circuit 7 successively;
Detector drives platform 3, for loading detector 2 and driving it to move;
Driver 4, for driving detector to drive platform 3 to move, and becomes multiple image on detector 2;
Memory cell 5, moves the multiple image of acquisition for real-time storage detector 2;
Image processing circuit 6, for reconstructing the multiple image in memory cell 5, obtains super-resolution image;
Image displaying circuit 7, for exporting the super-resolution image of reconstruct;
Described detector 2 is arranged on the focal plane of imaging device lens group 1, and is fixed on the side of detector driving platform 3 towards imaging device lens group 1; Described driver 4 and detector drive to adopt between platform 3 and are rigidly connected, and to make both movement operationally be consistent, and the size and Orientation of movement is random value.
Described detector 2 adopts charge coupled device ccd or complementary metal oxide semiconductors (CMOS) CMOS or charge injection device CID.
Described detector drives the movement of platform 3, be with to gather in the vertical plane of light carry out the randomized jitter of horizontal direction and vertical direction, and its in the horizontal direction jitter range be less than 0.23% of detector width size, the jitter range of in the vertical direction is less than 0.23% of detector length dimension.
Described detector drives the randomized jitter number of times of platform 3 identical with the number of image frames that described detector 2 obtains.
Described driver 4 adopts piezoelectric ceramic actuator.
Realize the formation method of gazing type super-resolution imaging, comprise the steps:
Step 1: according to required imaging effect, arranges the variation number of times instruction of driver drives part;
Step 2: utilize imaging lens head group 1 to gather light signal, obtain the light signal collected;
Step 3: according to the variation number of times instruction arranged, driver 4 drives detector 2 to make the randomized jitter of corresponding number of times respectively in the horizontal and vertical directions, jitter range is less than detector 2 length and wide 0.23%, the light signal collected imaging on detector 2, obtains the multiframe low-resolution image that frame number is corresponding to variation number of times;
Step 4: the multiframe low-resolution image utilizing the above-mentioned acquisition of memory cell 5 real-time storage;
Step 5: utilize the variation bayesian algorithm in image processing circuit 6, rebuilds the multiframe low-resolution image in memory cell 5, obtains corresponding super-resolution image;
Step 6: utilize image displaying circuit 7 to export above-mentioned through calculating and rebuilding the super-resolution image obtained.
The variation bayesian algorithm that above-mentioned steps five adopts comprises the steps:
Step 5.1 is set up reconstruction restricted model and is asked for posterior probability function P (x/y
k):
Rebuild restricted model y
k=DH
kc (s
k) x+n
k=B
k(s
k) x+n
k, wherein x represents the high-definition picture of scene, and its pixel count is PN; y
krepresent the low-resolution image obtained, its pixel count is N; P represents the raising of algorithm to image spatial resolution; K represents the number of image frames being not more than low-resolution image totalframes of acquisition; D is the down-sampled matrix of N × PN, H
kfor the kinematic matrix of PN × PN, C (s
k) be motion vector s
kthe kinematic matrix of the PN × PN produced, n
kbe the noise of N × 1, down-sampled, the fuzzy and anamorphic effect of imaging system can be combined into the sytem matrix B of a N × PN
k(s
k); The n of every frame low-resolution image
k, H
kand s
kcan not be identical, P (x/y can be obtained according to reconstruction restricted model
k), and then try to achieve its negative logarithm-lnP (y
k/ x);
Step 5.2 sets up prior-constrained model to ask for priori probability density function P (x):
Set up the feature space comprising image space dimensional information and directivity information be made up of the second-order partial differential coefficient of laplacian pyramid and gaussian pyramid, initial estimate is predicted as with laplacian pyramid, the estimation gradient priori comprising image high-frequency information is obtained by accelerating block matching method, namely obtain P (x), and then try to achieve its negative logarithm-lnP (x);
Step 5.3 is theoretical according to Bayesian MAP probability Estimation, in order to try to achieve optimum high-definition picture estimated value x, needing first to try to achieve and makes posterior probability P (x/y
k) get the numerical value of the x of maximum; Because y
kfor the known low-resolution image obtained, so P (y
k) be constant, thus high-definition picture maximum a posteriori probability is:
In formula
represent the super-resolution image of optimal estimation; The result of calculation utilizing above-mentioned two steps to draw, uses steepest descent method to try to achieve optimal estimation super resolution image in maximum a posteriori probability framework
, i.e. the final result of super-resolution image.
Compared with prior art, tool has the following advantages in the present invention:
1, the method that the present invention adopts multiple image to rebuild obtains output image, the method for relative general camera direct imaging, when using same probe, can obtain more high-resolution image.
2, imaging device of the present invention is to obtain multiple image by the mode of mobile detector, relative in prior art by moving lens or add optical element in the optical path mode for, technique is simple, cost is low and can reach and stare effect, improves the versatility of device.
3, in the present invention, the size and Orientation of detector shake displacement is random, in hinge structure, detector is fixing mobile, when kinematic parameter the unknown, can obtain super-resolution image, be applicable to the filming apparatus that self produces random file, as satellite.
4, in the present invention, the number of times of detector randomized jitter can by variation number of times instruction setting, in hinge structure, the mobile number of times of detector is a fixed value, variation number of times is more, the low-resolution image obtained is also corresponding to be increased, by the reconstruction in later stage, the resolution of the image finally exported is just higher.
5, in formation method of the present invention, apply variation bayesian algorithm when low-resolution image is rebuild, this algorithm introduces distribution and estimates, enhances the stability of algorithm to noise, effectively improves image resolution ratio.
Accompanying drawing explanation
Fig. 1 is structural representation of the present invention.
Fig. 2 is a kind of simple detector displacement mode schematic diagram in the present invention.
Fig. 3 is that in the present invention, detector moves four times at random, when displacement is half-pixel, and a kind of schematic diagram obtaining image that can select.
A kind of super-resolution reconstruction schematic diagram of Fig. 4 to be low-resolution image in the present invention be four frames.
Embodiment
In order to make object of the present invention, the technical problem solved and technical scheme more clear, below in conjunction with accompanying drawing, the present invention is described in further detail.
With reference to Fig. 1, the present invention includes imaging device lens group 1, detector 2, detector driving platform 3, driver 4, memory cell 5, image processing circuit 6 and image displaying circuit 7; In order to ensure that detector 2 successfully can receive image and drive platform 3 to be synchronized with the movement with detector, detector 2 is positioned in the light path of imaging device lens group 1 and in its focal plane, this detector 2 adopts charge coupled device ccd or complementary metal oxide semiconductors (CMOS) CMOS or charge injection device CID, before being fixed on detector driving platform 3; Driver 4 can adopt piezoelectric ceramic actuator, and be arranged at the side that detector drives platform 3, shell is fixedly connected with body, ensures that it itself does not produce relative movement in driving process; This driver 4 and detector platform 3 are rigidly connected, as bar connect, welding or bolt connect, in order to ensure being synchronized with the movement of detector platform 3 and detector 2; Detector 2 and memory cell 5, image processing circuit 6 and image displaying circuit 7 form circuit successively and are connected, for transmitting and process image.
Image-forming principle of the present invention is: according to required imaging effect, arranges the variation number of times instruction of driver drives part; Utilize imaging lens head group 1 to aim at collection that scene objects carries out light signal, obtains the light signal collected; When imaging device is started working, according to the variation number of times instruction of setting, the drive part of driver 4 impels detector to drive platform 3 to move, thus control the randomized jitter of detector 2 in the focal plane of lens group, detector 2 displacement is in the horizontal and vertical directions less than detector length and wide 0.23% respectively, and the size of this displacement is random value; Detector 2 obtains the frame number low-resolution image corresponding to the variation number of times of setting, frame number L represents, the every frame low-resolution image real-time storage got is in the middle of the memory cell 5 of camera, until the number of image frames stored in memory cell 5 reaches the L frame corresponding to the variation number of times instruction of setting; For after a complete L two field picture of scene capture, image digitization information transmission is reconstructed it to image processing circuit 6, obtains super-resolution image, finally by image displaying circuit 7, output display is carried out to the super-resolution image rebuild.
As follows to the principle of L two field picture reconstruction in above-mentioned image processing circuit 6:
The algorithm adopted in image processing circuit 6 is variation bayesian algorithm, the core of variation Bayesian Reconstruction algorithm, be utilize the priori such as the form of unknown probability density and the span of unknown parameter from the information of training sample itself, calculate posterior probability P (y
k/ x).
Rebuild restricted model:
y
k＝DH
kC(s
k)x+n
k＝B
k(s
k)x+n
k(1)
Wherein D is the down-sampled matrix of N × PN, H
kfor the kinematic matrix of PN × PN, C (s
k) be motion vector s
kthe kinematic matrix of the PN × PN produced, n
kbe the noise of N × 1, down-sampled, the fuzzy and anamorphic effect of imaging system can be combined into the sytem matrix B of a N × PN
k(s
k).The n of every frame low-resolution image
k, H
kand s
kcan not be identical.Wherein x represents the high-definition picture of scene, and its pixel count is PN; y
krepresent the low-resolution image obtained, its pixel count is N; P represents the raising of algorithm to image spatial resolution;
At known y
kcondition under, the posterior probability of x can be write as:
Wherein, P (y
k/ x) be x known when, observation y
kconditional probability density function; P (x) and P (y
k) represent x and y respectively
kprior probability.Theoretical according to Bayesian MAP probability Estimation, in order to try to achieve optimum estimation ultra-resolution rate image, must find and make posterior probability P (x/y
k) get the x of maximum.Because y
kknown, so P (y
k) be constant, logarithmic function is monotonically increasing function again, so maximum a-posteriori estimation can be formulated as follows:
represent the super-resolution image of optimal estimation.
From formula (3), as follows to the step of L two field picture reconstruction in above-mentioned image processing circuit 6:
Step 1: set up reconstruction restricted model and ask for-lnP (y
k/ x), rebuild restricted model such as formula shown in (1);
Step 2: set up prior-constrained model to ask for-lnP (x).Set up the feature space comprising image space dimensional information and directivity information be made up of the second-order partial differential coefficient of laplacian pyramid and gaussian pyramid, initial estimate is predicted as with laplacian pyramid, obtaining by accelerating block matching method the estimation gradient priori that contain image high-frequency information, namely obtaining-lnP (x);
Step 3: by result integration required by above-mentioned steps 1 and step 2 in maximum a posteriori probability framework, and use steepest descent method to try to achieve optimal estimation super resolution image x, i.e. the final result of super-resolution image.
Variation bayesian algorithm carries out Combined estimator under the super-resolution image of the unknown and kinematic parameter are placed on same framework, by super-resolution image and the kinematic parameter variance of probability distribution of the unknown, introduce certain uncertainty to estimation procedure, enhance the stability of algorithm to noise; And, adopt Bayesian Estimation method to carry out the point estimation of distribution estimation instead of traditional algorithm to unknown parameter in algorithm, effectively inhibit the expansion of evaluated error in algorithm, the resolution of image is improved.
With reference to Fig. 2, it is a kind of simple detector displacement mode schematic diagram in the present invention, the drive part of driver 4 impels detector to drive platform 3 to move, the detector 2 being in camera imaging lens group focal plane place is made to do randomized jitter in focal plane, when supposing that first time shakes, detector 2 center position coordinate is (0,0) the first two field picture y, is obtained
1, ensuing motion mode can be divided into following step:
1) with the first two field picture for reference, move horizontally detector 2 apart from c
2, vertical mobile detector 2 is apart from d
2, obtain the second frame low-resolution image y
2, its center position coordinate is (c
2, d
2), this image transmitting is temporarily preserved in memory cell 5.
2) with the first two field picture for reference, move horizontally detector 2 apart from c
3, vertical mobile detector 2 is apart from d
3, obtain the 3rd frame low-resolution image y
3, its center position coordinate is (c
3, d
3), this image transmitting is temporarily preserved in memory cell 5.
3) by that analogy, detector 2 from the shake of the 4th time to the L time, can obtain L ?3 two field pictures, in conjunction with three two field pictures of above-mentioned acquisition, accumulatively obtain L two field picture, real-time storage is in memory cell 5 successively.
4) will obtain in above-mentioned steps and be stored in the L two field picture of memory cell 5 li, being transferred to together in image processing circuit 6.
In order to algorithm for reconstructing that to make with L frame low-resolution image be data source reaches the reconstruction effect of super-resolution, the displacement between every two field picture of ensureing to select on horizontal and vertical direction is needed to be the distance c of sub-pix magnitude, i.e. movement
kand d
kit not the size of Integer Pel.L frame low-resolution image can be made to comprise the information of former scene different piece respectively by the randomized jitter controlling detector 2, but there is information redundancy each other in each low-resolution image, therefore can be rebuild by image processing circuit 6 in overlapping region, obtain super-resolution image, export finally by image displaying circuit 7.
With reference to Fig. 3, that in the present invention, detector moves four times at random, when displacement is half-pixel, possible a kind of schematic diagram obtaining image, Fig. 3 (a) represents the position of detector first time shake, Fig. 3 (b) represents that the position of detector shake relatively is for the first time shifted half-pixel left, Fig. 3 (c) represents that detector is upwards shifted half-pixel on the basis of Fig. 3 (b), and Fig. 3 (d) represents that detector is shifted half-pixel to the right on the basis of Fig. 3 (c).Super-resolution rebuilding process is carried out with reference to Fig. 4 to the four frame low-resolution images obtained, a kind of super-resolution reconstruction schematic diagram of to be low-resolution image in the present invention be four frames, in process of reconstruction, registration arrangement is carried out again to the pixel of each low-resolution image of Fig. 4 (a), obtain the super-resolution image that a frame resolution that Fig. 4 (b) shows significantly improves.
Claims (8)
1. a gazing type super-resolution imaging device, comprising:
Imaging device lens group (1), for gathering light signal, obtains the light signal collected;
Detector (2), for receiving the light signal collected, imaging thereon, and become image is exported to successively memory cell (5), image processing circuit (6) and image displaying circuit (7);
Detector drives platform (3), for loading detector (2) and driving it to move;
Driver (4), for driving detector to drive platform (3) mobile, and becomes multiple image on detector (2);
Memory cell (5), for the mobile multiple image obtained of real-time storage detector (2);
Image processing circuit (6), for reconstructing the multiple image in memory cell (5), obtains super-resolution image;
Image displaying circuit (7), for exporting the super-resolution image of reconstruct;
It is characterized in that: described detector (2) is arranged on the focal plane of imaging device lens group (1), and be fixed on the side of detector driving platform (3) towards imaging device lens group (1); Described driver (4) and detector drive to adopt between platform (3) and are rigidly connected, and to make both movement operationally be consistent, and the size and Orientation of movement is random value.
2. gazing type super-resolution imaging device according to claim 1, is characterized in that: described detector (2) adopts charge coupled device ccd or complementary metal oxide semiconductors (CMOS) CMOS or charge injection device CID.
3. gazing type super-resolution imaging device according to claim 1, it is characterized in that: described detector drives the movement of platform (3), be with to gather in the vertical plane of light carry out the randomized jitter of horizontal direction and vertical direction, and its in the horizontal direction jitter range be less than 0.23% of detector width size, the jitter range of in the vertical direction is less than 0.23% of detector length dimension.
4. gazing type super-resolution imaging device according to claim 3, is characterized in that: described detector drives the randomized jitter number of times of platform (3) identical with the number of image frames that described detector (2) obtains.
5. gazing type super-resolution imaging device according to claim 1, is characterized in that: described driver (4) adopts piezoelectric ceramic actuator.
6. utilize the device of claim 1 to carry out the method for gazing type super-resolution imaging, it is characterized in that: comprise the steps:
1) according to required imaging effect, the variation number of times instruction of driver is set;
2) utilize imaging lens head group to gather light signal, obtain the light signal collected;
3) according to the variation number of times instruction arranged, detector (2) is driven to make the randomized jitter of corresponding number of times respectively in the horizontal direction with vertical direction by driver (4), make the light signal collected in the upper imaging of detector (2), obtain the multiframe low-resolution image that frame number is identical with variation number of times;
4) the multiframe low-resolution image of the above-mentioned acquisition of memory cell (5) real-time storage is utilized;
5) utilize variation bayesian algorithm, rebuild the multiframe low-resolution image in memory cell (5), obtain corresponding super-resolution image;
6) the above-mentioned super-resolution image obtained through calculating and reconstruction is exported by image displaying circuit (7).
7. the method for gazing type super-resolution imaging according to claim 6, it is characterized in that: described step 3) in randomized jitter, its in the horizontal direction jitter range be less than 0.23% of detector width size, the jitter range of in the vertical direction is less than 0.23% of detector length dimension.
8. the method for gazing type super-resolution imaging according to claim 6, it is characterized in that: step 5) described in the variation bayesian algorithm utilized in image processing circuit (6), rebuild the multiframe low-resolution image in memory cell (5), carry out as follows:
5.1) set up reconstruction restricted model and ask for posterior probability function P (x/y
k):
Rebuild restricted model y
k=DH
kc (s
k) x+n
k=B
k(s
k) x+n
k, wherein x represents the high-definition picture of scene, and its pixel count is PN; y
krepresent the low-resolution image obtained, its pixel count is N; P represents the raising of algorithm to image spatial resolution; K represents the number of image frames being not more than low-resolution image totalframes of acquisition; D is the down-sampled matrix of N × PN, H
kfor the kinematic matrix of PN × PN, C (s
k) be motion vector s
kthe kinematic matrix of the PN × PN produced, n
kbe the noise of N × 1, down-sampled, the fuzzy and anamorphic effect of imaging system can be combined into the sytem matrix B of a N × PN
k(s
k); The n of every frame low-resolution image
k, H
kand s
kcan not be identical, P (x/y can be obtained according to reconstruction restricted model
k), and then try to achieve its negative logarithm-lnP (y
k/ x);
5.2) prior-constrained model is set up to ask for priori probability density function P (x):
Set up the feature space comprising image space dimensional information and directivity information be made up of the second-order partial differential coefficient of laplacian pyramid and gaussian pyramid, initial estimate is predicted as with laplacian pyramid, the estimation gradient priori comprising image high-frequency information is obtained by accelerating block matching method, namely obtain P (x), and then try to achieve its negative logarithm-lnP (x);
5.3) theoretical according to Bayesian MAP probability Estimation, in order to try to achieve optimum high-definition picture estimated value
need first to try to achieve and make posterior probability P (x/y
k) get the numerical value of the x of maximum; Because y
kfor the known low-resolution image obtained, so P (y
k) be constant, thus high-definition picture maximum a posteriori probability is:
In formula
represent the super-resolution image of optimal estimation; The result of calculation utilizing above-mentioned two steps to draw, uses steepest descent method to try to achieve optimal estimation super resolution image in maximum a posteriori probability framework
the i.e. final result of super-resolution image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410746361.0A CN104410789A (en) | 2014-12-08 | 2014-12-08 | Staring super-resolution imaging device and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410746361.0A CN104410789A (en) | 2014-12-08 | 2014-12-08 | Staring super-resolution imaging device and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104410789A true CN104410789A (en) | 2015-03-11 |
Family
ID=52648377
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410746361.0A Pending CN104410789A (en) | 2014-12-08 | 2014-12-08 | Staring super-resolution imaging device and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104410789A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104833307A (en) * | 2015-04-14 | 2015-08-12 | 中国电子科技集团公司第三十八研究所 | Method of high-frame-rate motioning object three-dimensional measurement |
| CN106525238A (en) * | 2016-10-27 | 2017-03-22 | 中国科学院光电研究院 | Spaceborne multispectral imaging system design method based on super-resolution reconstruction |
| CN106960416A (en) * | 2017-03-20 | 2017-07-18 | 武汉大学 | A kind of video satellite compression image super-resolution method of content complexity self adaptation |
| CN106982315A (en) * | 2017-04-01 | 2017-07-25 | 中国科学院上海应用物理研究所 | A kind of high-resolution imaging system |
| CN107194874A (en) * | 2017-05-26 | 2017-09-22 | 上海微小卫星工程中心 | Super-resolution imaging system and method based on bias image stabilization |
| CN110764085A (en) * | 2019-09-29 | 2020-02-07 | 西安电子科技大学 | Variational Bayes radar correlation imaging method combined with minimum mean square error estimation |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050019000A1 (en) * | 2003-06-27 | 2005-01-27 | In-Keon Lim | Method of restoring and reconstructing super-resolution image from low-resolution compressed image |
| CN101217625A (en) * | 2008-01-11 | 2008-07-09 | 清华大学 | Device and method of super-resolution imaging |
| CN101707670A (en) * | 2009-05-13 | 2010-05-12 | 西安电子科技大学 | Motion random exposure based super-resolution imaging system and method |
| CN101980291A (en) * | 2010-11-03 | 2011-02-23 | 天津大学 | Random micro-displacement-based super-resolution image reconstruction method |
| CN102006477A (en) * | 2010-11-25 | 2011-04-06 | 中兴通讯股份有限公司 | Image transmission method and system |
| US20120105690A1 (en) * | 2010-11-03 | 2012-05-03 | Sony Corporation | Camera system and imaging method using multiple lens and aperture units |
| CN102980664A (en) * | 2012-11-12 | 2013-03-20 | 北京航空航天大学 | Superpixel micro-scanning method and corresponding infrared super-resolution real-time imaging device |
-
2014
- 2014-12-08 CN CN201410746361.0A patent/CN104410789A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050019000A1 (en) * | 2003-06-27 | 2005-01-27 | In-Keon Lim | Method of restoring and reconstructing super-resolution image from low-resolution compressed image |
| CN101217625A (en) * | 2008-01-11 | 2008-07-09 | 清华大学 | Device and method of super-resolution imaging |
| CN101707670A (en) * | 2009-05-13 | 2010-05-12 | 西安电子科技大学 | Motion random exposure based super-resolution imaging system and method |
| CN101980291A (en) * | 2010-11-03 | 2011-02-23 | 天津大学 | Random micro-displacement-based super-resolution image reconstruction method |
| US20120105690A1 (en) * | 2010-11-03 | 2012-05-03 | Sony Corporation | Camera system and imaging method using multiple lens and aperture units |
| CN102006477A (en) * | 2010-11-25 | 2011-04-06 | 中兴通讯股份有限公司 | Image transmission method and system |
| CN102980664A (en) * | 2012-11-12 | 2013-03-20 | 北京航空航天大学 | Superpixel micro-scanning method and corresponding infrared super-resolution real-time imaging device |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104833307A (en) * | 2015-04-14 | 2015-08-12 | 中国电子科技集团公司第三十八研究所 | Method of high-frame-rate motioning object three-dimensional measurement |
| CN106525238A (en) * | 2016-10-27 | 2017-03-22 | 中国科学院光电研究院 | Spaceborne multispectral imaging system design method based on super-resolution reconstruction |
| CN106960416A (en) * | 2017-03-20 | 2017-07-18 | 武汉大学 | A kind of video satellite compression image super-resolution method of content complexity self adaptation |
| CN106960416B (en) * | 2017-03-20 | 2019-05-10 | 武汉大学 | A kind of video satellite that content complexity is adaptive compression image super-resolution method |
| CN106982315A (en) * | 2017-04-01 | 2017-07-25 | 中国科学院上海应用物理研究所 | A kind of high-resolution imaging system |
| CN107194874A (en) * | 2017-05-26 | 2017-09-22 | 上海微小卫星工程中心 | Super-resolution imaging system and method based on bias image stabilization |
| CN110764085A (en) * | 2019-09-29 | 2020-02-07 | 西安电子科技大学 | Variational Bayes radar correlation imaging method combined with minimum mean square error estimation |
| CN110764085B (en) * | 2019-09-29 | 2023-04-18 | 西安电子科技大学 | Variational Bayes radar correlation imaging method combined with minimum mean square error estimation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104410789A (en) | Staring super-resolution imaging device and method | |
| Chen et al. | Camera lens super-resolution | |
| CN103198523B (en) | A kind of three-dimensional non-rigid body reconstruction method based on many depth maps and system | |
| CN107959805B (en) | Light field video imaging system and method for processing video frequency based on Hybrid camera array | |
| Wang et al. | 360sd-net: 360 stereo depth estimation with learnable cost volume | |
| CN102868858B (en) | Image processing apparatus and image processing method | |
| CN102685378B (en) | Image pickup apparatus and image pickup optical system | |
| US9894252B2 (en) | Image processing apparatus, image pickup apparatus, image processing method, and storage medium for reducing noise of an image obtained by combining parallax images | |
| CN110596885B (en) | Scanning light field imaging system | |
| CN104113686B (en) | Camera device and its control method | |
| CN104365092A (en) | Method and apparatus for fusion of images | |
| CN103873758A (en) | Method, device and equipment for generating panorama in real time | |
| CN109087243A (en) | A kind of video super-resolution generation method generating confrontation network based on depth convolution | |
| CN206563985U (en) | 3-D imaging system | |
| CN102957864A (en) | Imaging device and control method thereof | |
| Zhao et al. | Super resolve dynamic scene from continuous spike streams | |
| JP2013042443A (en) | Image processing method, imaging apparatus, image processing apparatus, and image processing program | |
| CN108090869B (en) | On-satellite super-resolution reconstruction method based on area array CMOS optical camera | |
| CN103247020A (en) | Fisheye image spread method based on radial characteristics | |
| Liu et al. | A single frame and multi-frame joint network for 360-degree panorama video super-resolution | |
| CN104159119A (en) | Super-resolution reconstruction method and system for video images during real-time sharing playing | |
| CN104376547A (en) | Motion blurred image restoration method | |
| CN101707670B (en) | Motion random exposure based super-resolution imaging system and method | |
| Zhang et al. | Optical flow reusing for high-efficiency space-time video super resolution | |
| CN111369443A (en) | Zero-order learning super-resolution method for optical field cross-scale |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20150311 |
|
| WD01 | Invention patent application deemed withdrawn after publication |