CN102706449A - Two-channel remote sensing light spectrum imaging system based on compressed sensing and imaging method - Google Patents

Two-channel remote sensing light spectrum imaging system based on compressed sensing and imaging method Download PDF

Info

Publication number
CN102706449A
CN102706449A CN2012101727315A CN201210172731A CN102706449A CN 102706449 A CN102706449 A CN 102706449A CN 2012101727315 A CN2012101727315 A CN 2012101727315A CN 201210172731 A CN201210172731 A CN 201210172731A CN 102706449 A CN102706449 A CN 102706449A
Authority
CN
China
Prior art keywords
spectrum
image
information
dispersion element
coding
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.)
Granted
Application number
CN2012101727315A
Other languages
Chinese (zh)
Other versions
CN102706449B (en
Inventor
石光明
王立志
刘丹华
李国�
刘阳
高大化
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xidian University
Original Assignee
Xidian University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Xidian University filed Critical Xidian University
Priority to CN201210172731.5A priority Critical patent/CN102706449B/en
Publication of CN102706449A publication Critical patent/CN102706449A/en
Application granted granted Critical
Publication of CN102706449B publication Critical patent/CN102706449B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Image Processing (AREA)

Abstract

The invention discloses a two-channel remote sensing light spectrum imaging system based on compressed sensing, and an imaging method, and mainly solves the problems, in the prior art, that the utilization ratio of light spectrum information is low, the difficulty in manufacturing technique of a detector is high, high spatial resolution and high spectral resolution of spectrogram image can not be obtained at the same time. The imaging system comprises a beamsplitter module (1), a first observing passage module (2), a second observing passage module (3), and an image reconstructing and processing module (4). Original spectrogram images are divided into two paths of light beams which are the same in information and different in directions, the first observing passage module (2) and the second observing passage module (3) achieve complementary coding observation to the light spectrum image, observed results are output to the image reconstructing and processing module (4), a reconstruction model of the spectrogram image is built, and the original spectrogram images are reconstructed by using a nonlinear optimized method. The invention has the advantages that the utilization ratio of the light spectrum information is high, the manufacturing technique is simple, and the calculation complex rate is low, thereby being used for acquiring and reconstructing remote sensing spectrogram images.

Description

Binary channels remote sensing spectrum imaging system and formation method based on compressed sensing
Technical field
The invention belongs to technical field of image processing, further relate to a kind of optical spectrum encoded imaging system and formation method in the remote sensing field, can be used for realizing obtaining and reconstruct of remote sensing spectrum picture based on compressed sensing.
Background technology
Spectrum picture is defined in has increased the three-dimensional data cube that the spectrum dimension data is formed on the basis of traditional two-dimensional spatial domain, that is to say that spectrum picture is made up of the different spectral coverage image under the same field.Traditional panchromatic and coloured image can not satisfy people's application demand already far away, and all kinds of remote sensing light spectrum image-forming technology are by broad development.Several or tens wave bands of light spectrum image-forming techniques make use can be realized obtaining synchronously object space information, spectral information simultaneously to target imaging.Because its distinctive advantage that has imaging and spectrographic detection concurrently has been widely used in land ocean geography remote sensing, atmosphere environment supervision, military target is scouted, is kept watch on meteorological observation, a plurality of dual-use fields such as disaster prevention.
Because remote sensing light spectrum image-forming technology all has wide application potential in the civil and military field, the scientific research personnel is devoted to study various spectrum imaging systems and formation method always, but prior art meets with development bottleneck, mainly shows following two aspects:
First aspect, the spatial resolution of spectrum picture can be improved through the instantaneous field of view angle that reduces remote sensor; Spectral resolution can improve with the bandwidth that reduces each wave band through increasing wave band quantity.But under the certain condition of incident light energy, the narrow wave band of high-resolution spectra forms images and hangs down the contradiction between the reception of narrow-band radiated energy, causes the high spatial resolution of spectrum picture and spectral resolution not to obtain simultaneously.
Second aspect, spectrum picture are a kind of 3 d image datas, and its data volume is very huge.Particularly when the spectral resolution of image was improved, its data volume can sharply increase.In order to reduce the pressure of data transmission, prior art adopts the mode of image compression encoding to represent scene information with less data bit number always.Through data compression, a large amount of non-important data are abandoned, and the process of this high-speed sampling recompression causes the complexity of system to increase and wasted a large amount of sampling resources.
So very natural drawn a problem: can utilize other transformation spaces to describe signal; Set up the new signal description and the theoretical frame of processing; Make under the situation that guarantee information is not lost; Use the speed sampled signal far below the nyquist sampling theorem requirement, simultaneously can restoring signal fully again promptly be transformed into the sampling to information with signals sampling.
Since 2006, be born a kind of new compressed sensing of signal Processing field is theoretical, has attracted correlative study personnel's concern greatly.This theory is pointed out; When signal obtains, just data are suitably compressed, obtained and processing procedure, under the compressed sensing theoretical frame than traditional signal; Sampling rate no longer is decided by the bandwidth of signal; But being decided by the structure and the content of information in the signal, this makes the sampling and assessing the cost of sensor reduce greatly, and signal rejuvenation is the process of a computation optimization.The length of the signal X that note is sampled is N; Set sparse basic Ψ, make that promptly signal X is sparse on Ψ, the theoretical mathematical model of compressed sensing is observing matrix Φ who ties up with the incoherent M of Ψ * N of design; M＜N wherein, multiplying each other through Φ and X obtains the observation data Y of low dimension:
Y＝ΦX
Through finding the solution l 1Optimization problem under the norm is come reconstruct original signal X, and its process is:
min||Ψ TX|| 1?s.t.Y＝ΦX，
Wherein, Ψ is sparse base.
As everyone knows, the spectrum picture signal has compressibility, as long as select suitable sparse transform-based just can guarantee good degree of rarefication.At first can use one with the incoherent observing matrix of transform-based with on conversion gained higher-dimension signal projection to a lower dimensional space; Realize effective compression sampling of spectral information; So just can under certain spectral resolution condition, reduce the realization difficulty of camera, or under existence conditions, significantly put forward spectral resolution; Just can from these a spot of projections, reconstruct the original spectrum image through the solving-optimizing problem then, can prove that such projection has comprised the enough information of reconstruction signal with high probability.
According to above-mentioned theory, the scholar M.E.Gehm of Duke Univ USA, R.Johm; D.J.Brady, R.M.Willet and T.J.Schualz are at paper " Single-shot compressive spectral imaging with a dual-disperser architecture " OPTICS EXPRESS, Vol.15; No.21, pp.14013-14027 proposes to utilize random coded template and two dispersion elements in 2007; Realization is to the observation of spectrum picture; Reconstruct original image through the compressed sensing theory at last, the deficiency of this method is only to utilize single channel system to observe, and spectrum picture is losing half the effective information through meeting behind the coding templet; So just reduce the spectral information utilization factor of image, and then reduced the reconstruction accuracy of image.
Summary of the invention
The objective of the invention is to shortcoming and development bottleneck to above-mentioned prior art; A kind of binary channels remote sensing spectrum imaging system and formation method based on compressed sensing proposed; Realize the complementary encoding of spectrum picture, thereby improve the spectral information utilization factor and the reconstruction accuracy of image.
For realizing above-mentioned purpose, binary channels remote sensing spectrum imaging system of the present invention comprises the observation channel module; The image reconstruction process module, the observation channel module is observed spectrum picture, obtains observed image; The image reconstruction process module is carried out reconstruct to observed image; Obtain the original spectrum image, it is characterized in that, the observation channel module is divided into two; And the front end at these two observation passages is provided with splitter module; The incident beam of the spectrum picture of being gathered is divided into the light beam that two-way information is identical, direction is different through splitter module, and through the binary channels complementary encoding observation of the first observation channel module and the second observation channel module realization spectrum picture, observed result is exported to the reconstruct that the image reconstruction process module is carried out spectrum picture to this two-way light beam respectively.
The said first observation channel module comprises first lens combination, first dispersion element, first coding templet, second dispersion element, first planar array detector, and this first dispersion element is positioned at the rear end of first lens combination; And on the focal plane of first lens combination imaging, be used for the spectrum dimension information of translation spectrum picture, realize the chromatic dispersion of spectrum picture; This first coding templet is positioned at the rear end of first dispersion element, is used to realize the coding to spectrum picture, and this second dispersion element is positioned at the rear end of first coding templet; The spectrum dimension information that is used for reverse translation spectrum picture; To eliminate the chromatic dispersion effects of introducing by first dispersion element, realize the alignment again of spectrum dimension information, this first planar array detector is positioned at the second dispersion element rear end; Be used for observed image, obtain coding image information afterwards.
The said second observation channel module comprises second lens combination, the 3rd dispersion element, second coding templet, the 4th dispersion element, second planar array detector, and this second dispersion element is positioned at the rear end of second lens combination; And on the focal plane of second lens combination imaging, be used for the spectrum dimension information of translation spectrum picture, realize the chromatic dispersion of spectrum picture; This second coding templet is positioned at the rear end of the 3rd dispersion element, is used to realize the coding to spectrum picture, and the 4th dispersion element is positioned at the rear end of second coding templet; The spectrum dimension information that is used for reverse translation spectrum picture; To eliminate the chromatic dispersion effects of introducing by the 3rd dispersion element, realize the alignment again of spectrum dimension information, this second planar array detector is positioned at the 4th dispersion element rear end; Be used for observed image, obtain coding image information afterwards.
Described first coding templet and second coding templet, by the rectangle plane plate that printing opacity and lighttight grid are formed, the printing opacity grid is encoded to 1 to image, and light tight grid is encoded to 0 to image; Whether printing opacity is to set at random to each grid of first coding templet, realizes the random coded to each positional information of image; The light transmission state of the corresponding grid of the light transmission state of each grid of second coding templet and first coding templet is opposite, realizes the complementary encoding to each positional information of image.
For realizing above-mentioned purpose, binary channels remote sensing spectrum imaging method of the present invention comprises the steps:
(1) spectrum picture observation procedure:
(1a) establish original spectrum information matrix f 0Size be M * N * L, wherein M * N is the spectral information spatial resolution, L is a resolution between the spectrum of spectral information;
The spectral information of (1b) establishing arbitrarily any is f 0(m, n, k), and wherein m and n representation space dimension coordinate, k representes the spectrum dimension coordinate, 0≤m≤M-1 wherein, 0≤n≤N-1,0≤k≤L-1;
(1c) spectral information is divided into two-way in 1: 1 ratio, wherein the contained information f of the first via 11(m, n is k) with the second road contained information f 21(m, n, k) identical, that is:
f 11 ( m , n , k ) = 1 2 f 0 ( m , n , k ) ,
f 21 ( m , n , k ) = 1 2 f 0 ( m , n , k ) ;
(1d) with k pixel of information translation of k spectral coverage in the two-way spectral information, draw chromatic dispersion spectral information f afterwards 12(m, n, k) and f 22(m, n k) are respectively:
f 12 ( m , n , k ) = f 11 ( m - k , n , k ) = 1 2 f 0 ( m - k , n , k ) ,
f 22 ( m , n , k ) = f 21 ( m - k , n , k ) = 1 2 f 0 ( m - k , n , k ) ;
(1f) the two-way spectral information is encoded, coding function is respectively T 1(m, n) and T 2(m n), draws through the spectral information f after the coding 13(m, n, k) and f 23(m, n k) are respectively:
f 13 ( m , n , k ) = f 12 ( m , n , k ) T 1 ( m , n ) = 1 2 f 0 ( m - k , n , k ) T 1 ( m , n ) ,
f 23 ( m , n , k ) = f 22 ( m , n , k ) T 2 ( m , n ) = 1 2 f 0 ( m - k , n , k ) T 2 ( m , n ) ;
Wherein, T 1(m n) gets 0 or 1 randomly, T 2 ( m , n ) = 1 If T 1 ( m , n ) = 0 0 If T 1 ( m , n ) = 1 , To realize T 1(m, n) and T 2(m, the coding between n) is complementary;
(1g) with the reverse translation k pixel of information of k spectral coverage of two-way spectral information, the information of the same space position different spectral coverage of aliging again draws the spectral information f after the reverse translation 14(m, n, k) and f 24(m, n k) are respectively:
f 14 ( m , n , k ) = f 13 ( m + k , n , k ) = 1 2 f 0 ( m , n , k ) T 1 ( m + k , n ) ,
f 24 ( m , n , k ) = f 23 ( m + k , n , k ) = 1 2 f 0 ( m , n , k ) T 2 ( m + k , n ) ;
(1h) the two-way spectral information is made public, obtaining observed result is y 1(m, n) and y 2(m, n), wherein
y 1 ( m , n ) = Σ k f 14 ( m , n , k ) = 1 2 Σ k = 0 L - 1 f 0 ( m , n , k ) T 1 ( m + k , n ) ,
y 2 ( m , n ) = Σ k f 24 ( m , n , k ) = 1 2 Σ k = 0 L - 1 f 0 ( m , n , k ) T 2 ( m + k , n ) ;
Be designated as:
Y＝Hf；
Y={y wherein i(m, n) }, i=1,2 is the observed image matrix, H is a linear operator, the observation model of expression system, f is a part original spectrum information matrix;
(2) spectrum picture reconstruction step:
(2a) observed image matrix Y is delivered to image reconstruction processor;
(2b) setting sparse territory Ψ is DCT territory or wavelet field or Fourier domain, makes that spectrum picture is sparse under Ψ;
(2c) image reconstruction processor is utilized nonlinear optimization method reconstituting initial image f according to observed result Y and sparse territory Ψ.
The present invention compared with prior art has the following advantages:
First: the present invention has adopted twin-channel observation module; Realized the complementation observation of spectrum picture; Overcome in the existing imaging system spectral information utilization factor low with the low shortcoming of reconstruction accuracy, make the high and high advantage of reconstruction accuracy of tool spectral information utilization factor of the present invention;
Second: the light transmission state of the coding templet that the present invention adopts is set at random; Realized the random coded to spectrum picture, with respect to the mode of utilizing the exposure recompression earlier of high density detecting device in traditional imaging system, the present invention utilizes the low-density detector to make public; And the imaging compression is accomplished synchronously; Make that cost of the present invention is low, no compression artefacts, and system complexity is low;
The 3rd: the present invention has utilized the sparse property of spectrum picture, realizes image reconstruction through finding the solution nonlinear optimal problem, makes the present invention can obtain to have the spectrum picture of resolution between high spatial resolution and high spectrum simultaneously.
Description of drawings
Fig. 1 is the structured flowchart that the present invention is based on the binary channels remote sensing spectrum imaging system of coding perception;
Fig. 2 is the structured flowchart of the present invention's first observation channel module and the second observation channel module;
Fig. 3 is first coding templet and the second coding templet figure that uses among the present invention;
Fig. 4 is the process flow diagram that the present invention is based on the binary channels remote sensing spectrum imaging method of coding perception.
Embodiment
With reference to Fig. 1, the binary channels remote sensing spectrum imaging system based on compressed sensing of the present invention comprises that splitter module 1, the first observation channel module 2, second observation leads to module 3 and image reconstruction process module 4.Wherein splitter module 1 is positioned at the front end of the first observation passage 2 and the second observation passage 3; The first observation passage 2 is identical with the structure of the second observation passage 3; As shown in Figure 2; Fig. 2 (a) has provided the structure of the first observation passage 2, and Fig. 2 (b) has provided the structure of the second observation passage 3, and two input ends of image reconstruction process module 4 link to each other with the output terminal of the first observation passage and the second observation passage respectively.Splitter module 1 is divided into the light beam that two-way information is identical, direction is different with the incident beam of original spectrum image; This two-way light beam is respectively through the first observation channel module 2 and the second observation channel module 3; Realize the binary channels complementary encoding observation of spectrum picture; Observed result sends image reconstruction process module 4 to, and image reconstruction process module 4 is carried out the reconstruct of spectrum picture through nonlinear optimization method.
With reference to Fig. 2 (a), the first observation channel module 2 comprises: first lens combination 21, first dispersion element 22, first coding templet 23, second dispersion element 24 and first planar array detector 25.Wherein, first dispersion element 22 is positioned at the rear end of first lens combination 21, and on the focal plane of first lens combination, 21 imagings, is used for the spectrum dimension information of translation spectrum picture, realizes the chromatic dispersion of spectrum picture; The structure of first coding templet 23 is shown in Fig. 3 (a), and the rear end that it is positioned at first dispersion element 22 is used to realize the coding to spectrum picture; Second dispersion element 24 is positioned at the rear end of first coding templet 23; The spectrum dimension information that is used for reverse translation spectrum picture; The placement direction of its placement direction and first dispersion element 22 is opposite, to eliminate the chromatic dispersion effects of being introduced by first dispersion element, realizes the alignment again of spectrum dimension information; First planar array detector 25 is positioned at second dispersion element, 24 rear ends, is used for observed image, obtains coding image information afterwards.
Comprise with reference to Fig. 2 (b) second observation channel module 3: second lens combination 31, the 3rd dispersion element 32, second coding templet 33, the 4th dispersion element 34, second planar array detector 35.Wherein, second dispersion element 32 is positioned at the rear end of second lens combination 31, and on the focal plane of second lens combination, 31 imagings, is used for the spectrum dimension information of translation spectrum picture, realizes the chromatic dispersion of spectrum picture; Second coding templet 33 is shown in Fig. 3 (b), and the rear end that it is positioned at the 3rd dispersion element 32 is used to realize the coding to spectrum picture; The 4th dispersion element 34 is positioned at the rear end of second coding templet 33; The placement direction of its placement direction and the 3rd dispersion element 32 is opposite; Be used for the spectrum dimension information of reverse translation spectrum picture,, realize the alignment again of spectrum dimension information to eliminate the chromatic dispersion effects of introducing by the 3rd dispersion element; Second planar array detector 35 is positioned at the 4th dispersion element rear end, is used for observed image, obtains coding image information afterwards.
With reference to Fig. 3 (a) and Fig. 3 (b); The rectangle plane plate that first coding templet 32 and second coding templet 33 are made up of printing opacity and lighttight grid; Each grid size is identical; And with image slices vegetarian refreshments equal and opposite in direction, the printing opacity grid is encoded to 1 to image, and light tight grid is encoded to 0 to image; Each grid of first coding templet 23 whether set at random by printing opacity, realizes the random coded to each positional information of image; The light transmission state of the corresponding grid of the light transmission state of each grid of second coding templet 33 and first coding templet is opposite, realizes the complementary encoding to each positional information of image.
With reference to Fig. 4, for realizing above-mentioned purpose, the binary channels remote sensing spectrum imaging method based on compressed sensing of the present invention comprises spectrum picture observation and spectrum picture reconstruction step.
One, spectrum picture observation:
Step 1, initialization original spectrum information
If original spectrum information matrix f 0Size be M * N * L, wherein M * N is the spectral information spatial resolution, L is the spectral resolution of spectral information, promptly the spectral coverage number of spectral information is L;
Step 2, the spectral information of establishing arbitrarily any is f 0(m, n, k), and wherein m and n representation space dimension coordinate, k representes the spectrum dimension coordinate, 0≤m≤M-1 wherein, 0≤n≤N-1,0≤k≤L-1;
Step 3, spectral information are along separate routes
Spectral information is divided into two-way in 1: 1 ratio, wherein the contained spectral information f of the first via 11(m, n is k) with the second road contained spectral information f 21(m, n, k) identical, and be equal to Original spectrum information doubly, that is:
f 11 ( m , n , k ) = 1 2 f 0 ( m , n , k ) ,
f 21 ( m , n , k ) = 1 2 f 0 ( m , n , k ) .
Step 4, the spectral information translation
With first via spectral information f 11(m, n is k) with the second road spectral information f 21(m, n k) carry out linear translation respectively, are about to k pixel of information translation of k spectral coverage, draw chromatic dispersion two-way spectral information f afterwards 12(m, n, k) and f 22(m, n k) are respectively:
f 12 ( m , n , k ) = f 11 ( m - k , n , k ) = 1 2 f 0 ( m - k , n , k ) ,
f 22 ( m , n , k ) = f 21 ( m - k , n , k ) = 1 2 f 0 ( m - k , n , k ) .
Step 5, the spectral information coding
Set the first via coding function and the second road coding function and be respectively T 1(m, n) and T 2(m, n), respectively to the first via spectral information f after the chromatic dispersion 12(m, n is k) with the second road spectral information f 22(m, n k) encode, and draw coding two-way spectral information f afterwards 13(m, n, k) and f 23(m, n k) are respectively:
f 13 ( m , n , k ) = f 12 ( m , n , k ) T 1 ( m , n ) = 1 2 f 0 ( m - k , n , k ) T 1 ( m , n ) ,
f 23 ( m , n , k ) = f 22 ( m , n , k ) T 2 ( m , n ) = 1 2 f 0 ( m - k , n , k ) T 2 ( m , n ) ;
Wherein, T 1(m n) gets 0 or 1 randomly, T 2 ( m , n ) = 1 If T 1 ( m , n ) = 0 0 If T 1 ( m , n ) = 1 , To realize T 1(m, n) and T 2(m, the coding between n) is complementary.
Step 6, the reverse translation of spectral information
With the first via spectral information f after the coding 13(m, n is k) with the second road spectral information f 23(m, n k) carry out reverse translation, are about to the reverse translation k pixel of information of k spectral coverage, make the different spectral coverage information of the same space position to align again, draw the two-way spectral information f after the reverse translation 14(m, n, k) and f 24(m, n k) are respectively:
f 14 ( m , n , k ) = f 13 ( m + k , n , k ) = 1 2 f 0 ( m , n , k ) T 1 ( m + k , n ) ,
f 24 ( m , n , k ) = f 23 ( m + k , n , k ) = 1 2 f 0 ( m , n , k ) T 2 ( m + k , n ) .
Step 7 is to the first via spectral information f after the reverse translation 14(m, n is k) with the second road spectral information f 24(m, n k) make public respectively, obtain two-way observed result y 1(m, n) and y 2(m n) is respectively:
y 1 ( m , n ) = Σ k f 14 ( m , n , k ) = 1 2 Σ k = 0 L - 1 f 0 ( m , n , k ) T 1 ( m + k , n ) ,
y 2 ( m , n ) = Σ k f 24 ( m , n , k ) = 1 2 Σ k = 0 L - 1 f 0 ( m , n , k ) T 2 ( m + k , n ) .
Step 8 is with first via observed result y 1(m is n) with the second road observed result y 2(m n) merges into observed image matrix Y, i.e. Y={y i(m, n) }, i=1 wherein, 2, the linear observation model of initialization system is H, can draw:
Y＝Hf；
Wherein f is the original spectrum information matrix.
Two, spectrum picture reconstruct:
Step 1, Y is sent to image reconstruction processor with the observed image matrix.
Step 2, setting sparse transform domain Ψ is discrete sine transform territory or wavelet transformed domain or Fourier transform, makes that original spectrum image f is sparse on sparse transform domain Ψ, promptly makes the projection coefficient Ψ of original spectrum image f under sparse transform domain Ψ TMost numerical value are less than a certain specific threshold among the f, and this threshold value need be set through experiment, and different sparse transform domain corresponding threshold are different, and it is discrete cosine transform domain that this instance is set sparse transform domain Ψ, and setting threshold is 1, but the value of being not limited thereto.
Step 3, image reconstruction processor are utilized nonlinear optimization method reconstruct original spectrum image f according to observed result Y and sparse transform domain Ψ.
(3a) set the optimization aim function be min (|| Ψ TF|| 1), T representing matrix transposition wherein, || || 1Expression is got l to projection coefficient 1Norm, l is got in min () expression 1The minimum value of norm;
(3b) setting constraint condition is Hf=Y, and wherein Y is the observed image matrix, and H is the observation model of system, and f is the original spectrum image;
(3c) according to optimization aim function constraints, reconstruct original spectrum image f.

Claims (8)

1. the binary channels remote sensing spectrum imaging system based on compressed sensing comprises the observation channel module, the image reconstruction process module; The observation channel module is observed spectrum picture; Obtain observed image, the image reconstruction process module is carried out reconstruct to observed image, obtains the original spectrum image; It is characterized in that; The observation channel module is divided into two, and is provided with splitter module (1) at the front end of these two observation passages, and the incident beam of the spectrum picture of being gathered is divided into the light beam that two-way information is identical, direction is different through splitter module (1); Through the binary channels complementary encoding observation of first observation channel module (2) and second observation channel module (3) realization spectrum picture, export to image reconstruction process module (4) carry out the reconstruct of spectrum picture respectively by observed result for this two-way light beam.
2. the binary channels remote sensing spectrum imaging system based on compressed sensing according to claim 1 is characterized in that, the said first observation channel module (2); Comprise first lens combination (21), first dispersion element (22), first coding templet (23), second dispersion element (24), first planar array detector (25); This first dispersion element (22) is positioned at the rear end of first lens combination (21), and on the focal plane of first lens combination (21) imaging, is used for the spectrum dimension information of translation spectrum picture; Realize the chromatic dispersion of spectrum picture; This first coding templet (23) is positioned at the rear end of first dispersion element (22), is used to realize the coding to spectrum picture, and this second dispersion element (24) is positioned at the rear end of first coding templet (23); The spectrum dimension information that is used for reverse translation spectrum picture; To eliminate the chromatic dispersion effects of introducing by first dispersion element (22), realize the alignment again of spectrum dimension information, this first planar array detector (25) is positioned at second dispersion element (24) rear end; Be used for observed image, obtain coding image information afterwards.
3. the binary channels remote sensing spectrum imaging system based on compressed sensing according to claim 1 is characterized in that, the said second observation channel module (3); Comprise second lens combination (31), the 3rd dispersion element (32), second coding templet (33), the 4th dispersion element (34), second planar array detector (35); This second dispersion element (32) is positioned at the rear end of second lens combination (31), and on the focal plane of second lens combination (31) imaging, is used for the spectrum dimension information of translation spectrum picture; Realize the chromatic dispersion of spectrum picture; This second coding templet (33) is positioned at the rear end of the 3rd dispersion element (32), is used to realize the coding to spectrum picture, and the 4th dispersion element (34) is positioned at the rear end of second coding templet (33); The spectrum dimension information that is used for reverse translation spectrum picture; To eliminate the chromatic dispersion effects of introducing by the 3rd dispersion element (32), realize the alignment again of spectrum dimension information, this second planar array detector (35) is positioned at the 4th dispersion element (34) rear end; Be used for observed image, obtain coding image information afterwards.
4. according to claim 2 or 3 described binary channels remote sensing spectrum imaging systems based on compressed sensing; It is characterized in that; Said second dispersion element (24) is opposite with the placement direction of first dispersion element (22), and the 3rd dispersion element (32) is opposite with the placement direction of the 4th dispersion element (34).
5. according to claim 2 or 3 described binary channels remote sensing spectrum imaging systems based on compressed sensing; It is characterized in that first coding templet (23) and second coding templet (33); The rectangle plane plate of forming by printing opacity and lighttight grid; The printing opacity grid is encoded to 1 to image, and light tight grid is encoded to 0 to image; Each grid of first coding templet (23) whether set at random by printing opacity, realizes the random coded to each positional information of image; The light transmission state of the corresponding grid of the light transmission state of each grid of second coding templet (33) and first coding templet (23) is opposite, realizes the complementary encoding to each positional information of image.
6. the binary channels remote sensing spectrum imaging system based on compressed sensing according to claim 5; It is characterized in that first coding templet (23) is identical with light tight grid size with printing opacity grid in second coding templet (33), and the equal and opposite in direction of the size of each grid and image slices vegetarian refreshments.
7. binary channels remote sensing spectrum imaging method based on compressed sensing comprises:
(1) spectrum picture observation procedure:
(1a) establish original spectrum information matrix f 0Size be M * N * L, wherein M * N is the spectral information spatial resolution, L is the spectral resolution of spectral information;
The spectral information of (1b) establishing arbitrarily any is f 0(m, n, k), and wherein m and n representation space dimension coordinate, k representes the spectrum dimension coordinate, 0≤m≤M-1 wherein, 0≤n≤N-1,0≤k≤L-1;
(1c) spectral information is divided into two-way in 1: 1 ratio, wherein the contained information f of the first via 11(m, n is k) with the second road contained information f 21(m, n, k) identical, that is:
f 11 ( m , n , k ) = 1 2 f 0 ( m , n , k ) ,
f 21 ( m , n , k ) = 1 2 f 0 ( m , n , k ) ;
(1d) with k pixel of information translation of k spectral coverage in the two-way spectral information, draw chromatic dispersion spectral information f afterwards 12(m, n, k) and f 22(m, n k) are respectively:
f 12 ( m , n , k ) = f 11 ( m - k , n , k ) = 1 2 f 0 ( m - k , n , k ) ,
f 22 ( m , n , k ) = f 21 ( m - k , n , k ) = 1 2 f 0 ( m - k , n , k ) ;
(1f) the two-way spectral information is encoded, coding function is respectively T 1(m, n) and T 2(m n), draws through the spectral information f after the coding 13(m, n, k) and f 23(m, n k) are respectively:
f 13 ( m , n , k ) = f 12 ( m , n , k ) T 1 ( m , n ) = 1 2 f 0 ( m - k , n , k ) T 1 ( m , n ) ,
f 23 ( m , n , k ) = f 22 ( m , n , k ) T 2 ( m , n ) = 1 2 f 0 ( m - k , n , k ) T 2 ( m , n ) ;
Wherein, T 1(m n) gets 0 or 1 randomly, T 2 ( m , n ) = 1 If T 1 ( m , n ) = 0 0 If T 1 ( m , n ) = 1 , To realize T 1(m, n) and T 2(m, the coding between n) is complementary;
(1g) with the reverse translation k pixel of information of k spectral coverage of two-way spectral information, the information of the same space position different spectral coverage of aliging again draws the spectral information f after the reverse translation 14(m, n, k) and f 24(m, n k) are respectively:
f 14 ( m , n , k ) = f 13 ( m + k , n , k ) = 1 2 f 0 ( m , n , k ) T 1 ( m + k , n ) ,
f 24 ( m , n , k ) = f 23 ( m + k , n , k ) = 1 2 f 0 ( m , n , k ) T 2 ( m + k , n ) ;
(1h) the two-way spectral information is made public, obtaining observed result is y 1(m, n) and y 2(m, n), wherein
y 1 ( m , n ) = Σ k f 14 ( m , n , k ) = 1 2 Σ k = 0 L - 1 f 0 ( m , n , k ) T 1 ( m + k , n ) ,
y 2 ( m , n ) = Σ k f 24 ( m , n , k ) = 1 2 Σ k = 0 L - 1 f 0 ( m , n , k ) T 2 ( m + k , n ) ;
Be designated as:
Y＝Hf；
Y={y wherein i(m, n) }, i=1,2 is the observed image matrix, H is a linear operator, the observation model of expression system, f is the original spectrum information matrix;
(2) spectrum picture reconstruction step
(2a) observed image matrix Y is delivered to image reconstruction processor;
(2b) setting sparse territory Ψ is discrete cosine territory or wavelet field or Fourier domain, makes that spectrum picture is sparse under sparse territory Ψ;
(2c) image reconstruction processor is utilized nonlinear optimization method reconstituting initial image f according to observed result Y and sparse territory Ψ.
8. the binary channels remote sensing spectrum imaging method based on compressed sensing according to claim 7 is characterized in that, the described nonlinear optimization method reconstituting initial image that utilizes of step (2c) carries out as follows:
At first, set the optimization aim function be min (|| Ψ TF|| 1), T representing matrix transposition wherein, || || 1Expression is got l to projection coefficient 1Norm, l is got in min () expression 1The minimum value of norm;
Then, setting constraint condition is Hf=Y, and wherein Y is the observed image matrix, and H is the observation model of system, and f is the original spectrum image;
At last, according to optimization aim function constraints, reconstruct original spectrum image f.
CN201210172731.5A 2012-05-24 2012-05-24 Two-channel remote sensing light spectrum imaging system based on compressed sensing and imaging method Active CN102706449B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201210172731.5A CN102706449B (en) 2012-05-24 2012-05-24 Two-channel remote sensing light spectrum imaging system based on compressed sensing and imaging method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201210172731.5A CN102706449B (en) 2012-05-24 2012-05-24 Two-channel remote sensing light spectrum imaging system based on compressed sensing and imaging method

Publications (2)

Publication Number Publication Date
CN102706449A true CN102706449A (en) 2012-10-03
CN102706449B CN102706449B (en) 2014-03-19

Family

ID=46899399

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201210172731.5A Active CN102706449B (en) 2012-05-24 2012-05-24 Two-channel remote sensing light spectrum imaging system based on compressed sensing and imaging method

Country Status (1)

Country Link
CN (1) CN102706449B (en)

Cited By (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103400341A (en) * 2013-07-03 2013-11-20 西安电子科技大学 Method for recovering hyperspectral data by combining space and spectral domains based on compressive sensing
WO2013185589A1 (en) * 2012-06-13 2013-12-19 西安电子科技大学 Compression perception-based dual-channel multi-spectrum video imager and imaging method
CN103558160A (en) * 2013-10-21 2014-02-05 中国科学院遥感与数字地球研究所 Method and system for improving resolution ratio of spectral imaging space
CN103983355A (en) * 2014-05-27 2014-08-13 西安电子科技大学 Compressed spectrum imaging system and method based on panchromatic imaging
CN104111458A (en) * 2014-07-29 2014-10-22 西安电子科技大学 Method for compressed sensing synthetic aperture radar imaging based on dual sparse constraints
CN105139352A (en) * 2015-08-13 2015-12-09 西安电子科技大学 Quick spectrum video reconstruction method based on aliasing sampling of GPU
CN105758524A (en) * 2016-03-29 2016-07-13 西安电子科技大学 Spectrum camera based on all-pass single-template complementary sampling and imaging method
CN104200436B (en) * 2014-09-01 2017-01-25 西安电子科技大学 Multispectral image reconstruction method based on dual-tree complex wavelet transformation
CN106441577A (en) * 2016-09-27 2017-02-22 北京理工大学 Collaborative coding hyperspectral imaging system and image reconstruction method based on random projection
CN106998474A (en) * 2017-03-29 2017-08-01 南京大学 A kind of spectrum multichannel mixing compression transmitting method
CN107205103A (en) * 2017-04-14 2017-09-26 华东师范大学 Ultrahigh speed compression camera based on compressed sensing and streak camera principle
CN109444056A (en) * 2018-10-30 2019-03-08 浙江大学 A kind of underwater spectral reflectivity in-situ measurement device of binocular imaging formula and measurement method
CN109682476A (en) * 2019-02-01 2019-04-26 北京理工大学 A method of compression high light spectrum image-forming is carried out using adaptive coding aperture
CN112903103A (en) * 2021-01-22 2021-06-04 西安电子科技大学 Computed spectrum imaging system and method based on DMD and complementary all-pass
CN115100052A (en) * 2022-06-14 2022-09-23 北京理工大学 Wide-spectrum dual-channel compression imaging method and system

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7315371B2 (en) * 2004-01-23 2008-01-01 P&P Optica Inc. Multi-channel spectrum analyzer
CN101893552A (en) * 2010-07-06 2010-11-24 西安电子科技大学 Hyperspectral imager and imaging method based on compressive sensing

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7315371B2 (en) * 2004-01-23 2008-01-01 P&P Optica Inc. Multi-channel spectrum analyzer
CN101893552A (en) * 2010-07-06 2010-11-24 西安电子科技大学 Hyperspectral imager and imaging method based on compressive sensing

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
王伟伟等: "基于压缩感知的双通道SAR地面运动目标检测方法研究", 《电子与信息学报》, vol. 34, no. 3, 31 March 2012 (2012-03-31), pages 588 - 592 *

Cited By (26)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2013185589A1 (en) * 2012-06-13 2013-12-19 西安电子科技大学 Compression perception-based dual-channel multi-spectrum video imager and imaging method
US9851252B2 (en) 2012-06-13 2017-12-26 Xidian University Compressive sensing-based multispectral video imager with double channels and imaging method
CN103400341B (en) * 2013-07-03 2016-06-29 西安电子科技大学 Method based on the empty spectral domain integrated restoration high-spectral data of compressed sensing
CN103400341A (en) * 2013-07-03 2013-11-20 西安电子科技大学 Method for recovering hyperspectral data by combining space and spectral domains based on compressive sensing
CN103558160A (en) * 2013-10-21 2014-02-05 中国科学院遥感与数字地球研究所 Method and system for improving resolution ratio of spectral imaging space
CN103558160B (en) * 2013-10-21 2015-10-28 中国科学院遥感与数字地球研究所 A kind of method and system improving light spectrum image-forming spatial resolution
CN103983355A (en) * 2014-05-27 2014-08-13 西安电子科技大学 Compressed spectrum imaging system and method based on panchromatic imaging
CN103983355B (en) * 2014-05-27 2016-01-20 西安电子科技大学 Based on compressed spectrum imaging system and the formation method of full color imaging
CN104111458A (en) * 2014-07-29 2014-10-22 西安电子科技大学 Method for compressed sensing synthetic aperture radar imaging based on dual sparse constraints
CN104111458B (en) * 2014-07-29 2016-08-24 西安电子科技大学 Compressed sensing synthetic aperture radar image-forming method based on dual sparse constraint
CN104200436B (en) * 2014-09-01 2017-01-25 西安电子科技大学 Multispectral image reconstruction method based on dual-tree complex wavelet transformation
CN105139352A (en) * 2015-08-13 2015-12-09 西安电子科技大学 Quick spectrum video reconstruction method based on aliasing sampling of GPU
CN105758524A (en) * 2016-03-29 2016-07-13 西安电子科技大学 Spectrum camera based on all-pass single-template complementary sampling and imaging method
CN106441577A (en) * 2016-09-27 2017-02-22 北京理工大学 Collaborative coding hyperspectral imaging system and image reconstruction method based on random projection
CN106441577B (en) * 2016-09-27 2018-01-09 北京理工大学 Collaboration coding Hyperspectral imager and image reconstructing method based on accidental projection
CN106998474A (en) * 2017-03-29 2017-08-01 南京大学 A kind of spectrum multichannel mixing compression transmitting method
CN106998474B (en) * 2017-03-29 2019-10-18 南京大学 A kind of spectrum multichannel mixing compression transmitting method
CN107205103A (en) * 2017-04-14 2017-09-26 华东师范大学 Ultrahigh speed compression camera based on compressed sensing and streak camera principle
CN107205103B (en) * 2017-04-14 2020-02-14 华东师范大学 Ultra-high speed compression photographic device based on compressed sensing and stripe camera principle
CN109444056A (en) * 2018-10-30 2019-03-08 浙江大学 A kind of underwater spectral reflectivity in-situ measurement device of binocular imaging formula and measurement method
CN109444056B (en) * 2018-10-30 2024-03-01 浙江大学 Binocular imaging type underwater spectral reflectivity in-situ measurement device and measurement method
CN109682476A (en) * 2019-02-01 2019-04-26 北京理工大学 A method of compression high light spectrum image-forming is carried out using adaptive coding aperture
CN109682476B (en) * 2019-02-01 2020-06-19 北京理工大学 Method for performing compressed hyperspectral imaging by using self-adaptive coding aperture
CN112903103A (en) * 2021-01-22 2021-06-04 西安电子科技大学 Computed spectrum imaging system and method based on DMD and complementary all-pass
CN115100052A (en) * 2022-06-14 2022-09-23 北京理工大学 Wide-spectrum dual-channel compression imaging method and system
CN115100052B (en) * 2022-06-14 2024-06-07 北京理工大学 Wide-spectrum dual-channel compression imaging method and system

Also Published As

Publication number Publication date
CN102706449B (en) 2014-03-19

Similar Documents

Publication Publication Date Title
CN102706449B (en) Two-channel remote sensing light spectrum imaging system based on compressed sensing and imaging method
CN103983355B (en) Based on compressed spectrum imaging system and the formation method of full color imaging
CN102706450B (en) Dual-channel multispectral video imaging device and imaging method based on compressive sensing
CN101893552B (en) Hyperspectral imager and imaging method based on compressive sensing
CN103428500B (en) A kind of super-resolution big visual field infrared imaging method
CN104154998B (en) A kind of reconstructing method of the calculating multispectral imaging collection of illustrative plates based on compressed sensing
CN108955882B (en) Three-dimensional data reconstruction method based on liquid crystal hyperspectral calculation imaging system
CN102818631B (en) Compression spectral imaging system based on random wavefront phase modulation
EP3182080B1 (en) Compressed sensing broadband hyperspectral imaging system based on random grating
CN103592026B (en) Time flight imaging spectrum system and method based on compressed sensing and coding transformation
CN101975947B (en) Two-dimensional radial imaging method for mirror image synthetic aperture
CN204788661U (en) Calculate many spectral imaging system based on compressed sensing
CN104793343A (en) Three-channel and single-Wollaston prism polarization imaging device and polarization information detecting method
CN105954750B (en) The non-sparse scene imaging method of stripmap synthetic aperture radar based on compressed sensing
CN105758524A (en) Spectrum camera based on all-pass single-template complementary sampling and imaging method
CN103913229B (en) Coding template multi-target active imaging spectrum system and method
CN114387164A (en) Terahertz single-pixel super-resolution imaging method and system
CN105182359A (en) Satellite-borne Lidar hyperchaotic compressed sensing high-spatial-resolution imaging method
Xiao et al. A compressed sensing approach for enhancing infrared imaging resolution
CN106441577B (en) Collaboration coding Hyperspectral imager and image reconstructing method based on accidental projection
CN105021280A (en) Panchromatic-and-compressed-spectral-imaging-based image super-resolution system and method
CN108663118A (en) Infrared broadband EO-1 hyperion calculates imaging devices and methods therefor
CN114638758A (en) Method for restoring and reconstructing damaged image based on spectral imaging technology of compressed sensing
CN109682476A (en) A method of compression high light spectrum image-forming is carried out using adaptive coding aperture
CN110736539B (en) Gaze type spectral imaging system based on compressed sensing

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C53 Correction of patent of invention or patent application
CB03 Change of inventor or designer information

Inventor after: Liu Danhua

Inventor after: Li Guo

Inventor after: Shi Guangming

Inventor after: Gao Dahua

Inventor after: Wang Lizhi

Inventor after: Liu Yang

Inventor before: Shi Guangming

Inventor before: Wang Lizhi

Inventor before: Liu Danhua

Inventor before: Li Guo

Inventor before: Liu Yang

Inventor before: Gao Dahua

COR Change of bibliographic data

Free format text: CORRECT: INVENTOR; FROM: SHI GUANGMING WANG LIZHI LIU DANHUA LI GUO LIU YANG GAO DAHUA TO: LIU DANHUA LI GUO SHI GUANGMING GAO DAHUA WANG LIZHI LIU YANG

GR01 Patent grant
GR01 Patent grant