CN110244299A - A kind of distributed method that the SAR image based on ADMM is restored - Google Patents
A kind of distributed method that the SAR image based on ADMM is restored Download PDFInfo
- Publication number
- CN110244299A CN110244299A CN201910544976.8A CN201910544976A CN110244299A CN 110244299 A CN110244299 A CN 110244299A CN 201910544976 A CN201910544976 A CN 201910544976A CN 110244299 A CN110244299 A CN 110244299A
- Authority
- CN
- China
- Prior art keywords
- sar image
- distributed
- admm
- algorithm
- restored
- 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
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/89—Radar or analogous systems specially adapted for specific applications for mapping or imaging
- G01S13/90—Radar or analogous systems specially adapted for specific applications for mapping or imaging using synthetic aperture techniques, e.g. synthetic aperture radar [SAR] techniques
- G01S13/9004—SAR image acquisition techniques
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/89—Radar or analogous systems specially adapted for specific applications for mapping or imaging
- G01S13/90—Radar or analogous systems specially adapted for specific applications for mapping or imaging using synthetic aperture techniques, e.g. synthetic aperture radar [SAR] techniques
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
Landscapes
- Engineering & Computer Science (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- General Physics & Mathematics (AREA)
- Electromagnetism (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The distributed method that a kind of SAR image based on ADMM provided by the invention is restored, comprising the following steps: step 1, construct observing matrix;Step 2, SAR image sparse information Restoration model is constructed according to the observing matrix that step 1 obtains;Step 3, Distributed Problem Solving Algorithm of the strategy design based on ADMM algorithm SAR image sparse information Restoration model obtained in step 2 being grouped according to variable;Step 4, the Distributed Problem Solving Algorithm based on ADMM algorithm obtained using the design in Spark platform building step 3, finally obtains the SAR image restored;The present invention establishes compressed sensing model using information theory principle and machine learning distributed algorithm, sparse reconstruction target is solved by distributed ADMM algorithm, the present invention combines the compressed sensing of SAR image with machine learning distributed algorithm, the on-board SAR image being able to solve under any motion state and the Problems of Reconstruction to large scale scene have very important theoretical and practical significance.
Description
Technical field
The present invention relates to electronics industry Radar Technology field, in particular to what a kind of SAR image based on ADMM was restored divides
Cloth method.
Background technique
Synthetic aperture radar (Synthetic Aperture Radar, abbreviation SAR) technology has round-the-clock, round-the-clock etc.
Feature, it is widely used in the fields such as environmental protection, disaster monitoring, oceanographic observation.Currently, synthetic aperture radar technique is
Developed towards broader mapping band with the direction for obtaining high-definition picture, thus to data sampling, transmission, processing etc.
It is required that higher and higher.Sparse synthetic aperture radar image-forming is a kind of by sparse information processing method and synthetic aperture radar image-forming skill
The new imaging pattern that art combines, the technology pass through compression sampling and utilize atural object scene sparsity, are being far below Nyquist
Exact Reconstruction target under hits required by sampling thheorem.Compressed sensing (compression sensing, abbreviation CS) method
It is a kind of typical sparse signal processing method that developed recently gets up, it can realize sparse letter from serious lack sampling data
Number accurate or approximate reconstruction.In recent years, it reducing radar sampling about the compressed sensing technology of SAR image, making up radar number
Great potential is shown according to self-defect, improvement radar imagery quality etc., and achieves certain achievement, but this method pair
It still can not be rebuild in large scene SAR image.Since big data Distributed Computing Platform is increasingly mature, it is this technology of solution
Problem provides new approach.Extensive SAR image compressed sensing problem is practical problem urgently to be resolved, utilizes big data point
Cloth computing technique realizes the reconstruction to extensive SAR scene image, has very important theoretical and practical significance.
Summary of the invention
The distributed method that the SAR image based on ADMM that the purpose of the present invention is to provide a kind of is restored, solves existing
Conventional method the problem of can not being rebuild for large scene SAR image.
In order to achieve the above object, the technical solution adopted by the present invention is that:
The distributed method that a kind of SAR image based on ADMM provided by the invention is restored, comprising the following steps:
Step 1, observing matrix is constructed;
Step 2, SAR image sparse information Restoration model is constructed according to the observing matrix that step 1 obtains;
Step 3, strategy design base SAR image sparse information Restoration model obtained in step 2 being grouped according to variable
In the Distributed Problem Solving Algorithm of ADMM algorithm;
Step 4, it is calculated using distributed solve based on ADMM algorithm that the design in Spark platform building step 3 obtains
Method finally obtains the SAR image restored.
Preferably, in step 1, the construction method of observing matrix is: by two-dimentional sampled echo and two dimension image to be reconstructed to
Quantization.
Preferably, in step 2, the mathematic(al) representation of SAR image sparse information Restoration model are as follows:
S.t.y=Ax
Wherein, x ∈ RnIndicate reconstruction signal, A ∈ Rm×nFor the observing matrix of construction, y ∈ RmIndicate the signal observed;|
|·||1Indicate 1- norm.
Preferably, in step 3, to the plan that SAR image sparse information Restoration model is grouped according to variable obtained in step 2
The Distributed Problem Solving Algorithm based on ADMM algorithm is slightly designed, is specifically included:
S1, the SAR image that the SAR image sparse information Restoration model according to obtained in step 2 constructs variable grouping are sparse
Information recovering distributed model;
S2 restores distributed model to SAR image sparse information using distributed ADMM algorithm and solves, is based on
The Distributed Problem Solving Algorithm of ADMM algorithm.
Preferably, in S1, the SAR image sparse information Restoration model according to obtained in step 2 constructs the SAR of variable grouping
Image sparse Information recovering distributed model, specifically includes:
Setting SAR image sparse information Restoration model includes N number of independent subproblem, and each independent subproblem is carried out
Parallelization solves, and the SAR image sparse information for obtaining variable grouping restores distributed model;SAR image sparse information restores to divide
The mathematic(al) representation of cloth model are as follows:
Wherein, A=[A1;A2;…;AN],λ is regularization parameter.
Preferably, in S2, the expression formula of the Distributed Problem Solving Algorithm based on ADMM algorithm are as follows:
Wherein, Sλ(a) it is expressed as soft threshold operator to act on the vector that each component of a obtains, concrete form
It is as follows:
giIndicate gradient,τ, β, γ are
Initiation parameter, pnFor iteration m ultiple.
Preferably, in step 4, the distribution based on ADMM algorithm that is obtained using the design in Spark platform building step 3
Formula derivation algorithm, specifically includes:
S1, initiation parameter, including regularization parameter λ, τ, beta, gamma step-length stepsize and iteration precision ε > 0, simultaneously
Provide initial point x0,p0,res0;
S2 utilizes the distributed multiplication computation residual error res=Ax-b of matrix and vector;
S3 utilizes the distributed multiplication computation gradient g=A of matrix transposition and vector*(Ax-b-p/ β), wherein p is algorithm
Multiplier;
S4 is updated x=soft (x-stepsizeg) using threshold operator;
S5, if | | xk+1-xk| |≤ε, then iteration stopping, exports xk, wherein xkFor the SAR image restored;Otherwise turn
To S2.
Compared with prior art, the beneficial effects of the present invention are:
The distributed method that a kind of SAR image based on ADMM provided by the invention is restored, solves existing tradition side
The problem of method can not rebuild large scene SAR image, specifically:
Firstly, since sparse synthetic aperture radar image-forming is largely handled in time domain, the calculating of existing method is complicated
Degree and memory consumption are excessively high, restore to bring difficulty to the imaging of spaceborne or airborne large scene.The present invention utilizes Spark big data platform
The distributed method that the SAR image based on ADMM is restored is realized, solves existing conventional method for large scene SAR image
The problem of can not rebuilding.The invention is that earth remote sensing field brings huge potentiality, and on the one hand it reduces radar sampling rate,
By the hardware store and processing of high radar memory transfer to ground, on the other hand the technology can make up for it radar data itself
Noise and defect, played important function in terms of improving radar imagery quality;
Secondly, ADMM algorithm is converted to distributed algorithm by the strategy according to variable grouping, ADMM algorithm has easily distribution
The characteristics of change, decomposability, it can effectively solve the problem that the bigger compressed sensing of data scale, signal processing and control, picture are deposited
The problems such as storage is with rebuilding;
Finally, the distributed model and distribution ADMM method for solving of variable grouping are proposed for large-scale data, benefit
With Spark big data platform it is scalable, based on memory the features such as, interative computation efficiency far be higher than MapReduce etc. its
His big data platform, has efficiently agreed with ADMM algorithm iteration frame, take full advantage of the operability of spark platform calculating with
High efficiency.
In conclusion the present invention establishes compressed sensing model using information theory principle and machine learning distributed algorithm, lead to
It crosses distribution ADMM algorithm and solves sparse reconstruction target.The compressed sensing problem of SAR image is the hot issue of big data field,
It is combined with machine learning distributed algorithm, the on-board SAR image that is able to solve under any motion state and to extensive field
The Problems of Reconstruction of scape has very important theoretical and practical significance, and the method for the present invention has promotes and applies scene well.
Detailed description of the invention
Fig. 1 is overall flow schematic diagram of the present invention;
Fig. 2 is the frame flow diagram that the present invention solves sparse model using ADMM algorithm;
Fig. 3 is the MapReduce schematic diagram of ADMM distributed algorithm of the present invention;
Fig. 4 is truthful data effect picture of the invention.
Specific embodiment
With reference to the accompanying drawing, the present invention is described in more detail.
As shown in Figure 1, the present invention provides a kind of distributed method that effectively SAR image based on ADMM is restored, benefit
It is established with variable grouping strategy and is based on compressed sensing imaging radar model, and pass through distribution using Spark big data platform
ADMM method solves sparse model and correctly rebuilds target, specifically includes the following steps:
Step 1, observing matrix is constructed: by two-dimentional sampled echo and two dimension image vector to be reconstructed;
Step 2, SAR image sparse information Restoration model is constructed according to the observing matrix that step 1 obtains, wherein SAR image
The mathematic(al) representation of sparse information Restoration model are as follows:
S.t.y=Ax
Wherein, x ∈ RnIndicate reconstruction signal, A ∈ Rm×nFor the observing matrix of construction, y ∈ RmIndicate the signal observed.|
|·||1Indicate 1- norm.
Step 3, strategy design base SAR image sparse information Restoration model obtained in step 2 being grouped according to variable
In the Distributed Problem Solving Algorithm of ADMM algorithm;
As shown in Fig. 2, specifically:
In S1, the SAR image that the SAR image sparse information Restoration model according to obtained in step 2 constructs variable grouping is dilute
Information recovering distributed model is dredged, is specifically included:
Setting SAR image sparse information Restoration model includes N number of independent subproblem, and each independent subproblem is carried out
Parallelization solves, and the SAR image sparse information for obtaining variable grouping restores distributed model.SAR image sparse information restores to divide
The mathematic(al) representation of cloth model are as follows:
Wherein, A=[A1;A2;…;AN],λ is regularization parameter.
S2 restores distributed model to SAR image sparse information using distributed ADMM algorithm and solves, solves expression
Formula are as follows:
Wherein, soft is soft-threshold algorithm, Sλ(a) it is expressed as soft threshold operator and acts on each component of a obtaining
Vector on, concrete form is as follows:
giIndicate gradient,τ, β, γ are
Initiation parameter, pnFor iteration m ultiple.
Step 4, it is calculated using distributed solve based on ADMM algorithm that the design in Spark platform building step 3 obtains
Method, and correct the punishment parameter λ of iterative algorithm with convergence rate according to convergence and carry out data in Spark platform and delay
It deposits, the tuning in terms of checkpointing and parameter configuration, finally obtains the SAR image restored.
As shown in figure 3, being asked using the distribution based on ADMM algorithm that the design in Spark platform building step 3 obtains
Resolving Algorithm specifically includes:
S1, initiation parameter, including regularization parameter λ, τ, beta, gamma, step-length stepsize and iteration precision ε > 0, simultaneously
Provide initial point x0,p0,res0;
S2 utilizes the distributed multiplication computation residual error res=Ax-b of matrix and vector;
S3 utilizes the distributed multiplication computation gradient g=A of matrix transposition and vector*(Ax-b-p/ β), wherein p is algorithm
Multiplier;
S4 is updated x=soft (x-stepsizeg) using threshold operator;
S5, if | | xk+1-xk| |≤ε, then iteration stopping, exports xk, xkFor the SAR image restored;Otherwise S2 is gone to.
It can be seen via above technical scheme that disclosed by the embodiments of the present invention is a kind of SAR image recovery based on ADMM
Distributed method, establish compressed sensing model using information theory principle and machine learning distributed algorithm, pass through distribution
ADMM algorithm solves sparse reconstruction target.The compressed sensing problem of SAR image is the hot issue of big data field, with engineering
It practises distributed algorithm to combine, the on-board SAR image being able to solve under any motion state and the reconstruction to large scale scene
Problem has very important theoretical and practical significance, and the method for the present invention has promotes and applies scene well.
The principle of the present invention:
In the prior art, image sparse information Restoration model, image sparse are constructed according to the observing matrix that step 1 obtains
The expression formula of Information recovering model:
Wherein, x ∈ RnIndicate reconstruction signal, A ∈ Rm×nFor the observing matrix of construction, y ∈ RmIndicate the signal observed.
Due to L0Problem is np hard problem, is usually translated into Lq(0 q≤1 <) problem is solved:
The theoretical research of compressed sensing shows to meet RIP in observing matrix, can be by asking when the properties such as incoherent
Solve LqProblem equivalent solves L0Problem, and q is smaller, LpSolution closer to L0Solution, but corresponding to solve difficulty bigger.With
High-resolution image request is obtained, the increase for surveying and drawing bandwidth directly results in calculation matrix and becomes larger, thus nothing under the conditions of single machine
Method solves, therefore we acquire true or approximate solution using big data platform combination distributed algorithm.
Work as q=0, when 1/2,2/3,1, the above problem there are analytic solutions, it is contemplated that the sparsity and solving speed of solution, Wo Menxuan
With soft-threshold operator (q=1), derives be equally applicable to complex number space above, it is possible to be applied directly to synthetic aperture radar
In imaging.
For extensive problem, traditional threshold value alternative manner convergence rate is excessively slow, and the application uses ADMM algorithm
The problem is solved, the specific steps are as follows:
Regularization parameter λ is added, model becomes following form:
Regularization parameter λ is used to balance the sparsity of approximation capability and solution.
Multiplier p is introduced, above-mentioned model is solved using ADMM algorithm, obtains following computation model:
Wherein, soft is soft-threshold operator, and concrete form is
softλ(a)=sign (a) max 0, | and a |-λ }
Since data volume is very big, using big data distributed algorithm solve, due to observing matrix A be by column-generation,
Variable is grouped by consideration, and model is as follows:
minimize f1(x1)+…+fN(xN)
subject to A1x1+…+ANxN=c,
x1∈χ1,…,xN∈χN.
It is assumed that fi(xi) it is convex function, then the Augmented Lagrangian Functions of the above problem are
WhereinFor Lagrange multiplier, ρ is penalty parameter;
Thus parallelization update is obtained
Variable is grouped by the present invention, considers compressed sensing model:
Distributed model is solved to obtain such as lower frame:
Wherein, giFor gradient,
It is tested using Spark big data platform, Spark platform is emerging big data engine, is mainly characterized by mentioning
It has supplied the distributed memory of a data set abstract, has been calculated based on memory, provide better support for iterative data processing.
Diameter radar image records the echo information of wave band, is recorded with binary complex format, based on each
The corresponding amplitude of the convertible extraction of the complex data of pixel and phase information.The real part of observing matrix is deposited by column respectively with imaginary part
Storage, since the plural packet committed memory that Spark platform carries internally will have very big requirement greatly more than type double precision, this
Application with type double precision writes the complex operation packet with whole complex operations to be calculated the requirement to reduce to memory,
Realize the distributed multiplication of the distributed multiplication of matrix and vector and the transposition of matrix and vector respectively simultaneously.
Embodiment
Specific numerical operation example and true SAR image example is given below to illustrate the validity of algorithm, it is all
Experiment realized on Spark platform using Scala language, the cluster environment totally 28 nodes, respectively iNode1-
INode28, the free memory of each node are 256G, and 12 cores can save as 4.6T with always interior, and nucleus number is 324 cores, and Driver is available
300G is inside saved as, Exector free memory is 256G.
Table one: numerical result
Data dimension | Size of data | Iterative steps | Precision | Time used |
(500000*1500000) | 55.9G | 100 | 1.24e-4 | 138s |
(512*256)*(512*512) | 234.0G | 100 | 9.00e-5 | 234s |
(512*512)*(1024*512) | 937.5G | 100 | 8.42e-5 | 1309s |
(3000000*6000000) | 1341.1G | 100 | 8.30e-5 | 1674s |
It is demonstrated experimentally that algorithm is feasible under big data Spark platform, and arithmetic speed is quickly.
The validity that true SAR image example carrys out verification algorithm is given below, the following figure is the synthetic aperture of RADARSAT-1
Radar image, for this Image Acquisition in Green George time 02:03:50 to 02:04:05 on June 16th, 2002, rail lift number is #
34522, fine pattern wave beam 2, image size is 1024 × 512, and obtaining perception matrix using 30% sample rate is 571392
× (1024*512) and be complex data, stores nearly 3.6T using type double precision, wherein parameter setting τ=10/ | | y | |, η
=10/ | | y | |, λ=0.2, γ=1;Fig. 4 be run 180 steps obtain as a result, at this time
RNorm=0.003371 (rNorm=| | x-xold)/| | x | |)
In above-described specific example, the purpose of the present invention, technical scheme and beneficial effects have been carried out further
It is described in detail, it should be understood that the above description is only a specific example of the present invention, is not intended to restrict the invention,
All within the spirits and principles of the present invention, any modification, equivalent substitution, improvement and etc. done should all include of the invention
Within protection scope.
Claims (7)
1. a kind of distributed method that effectively SAR image based on ADMM is restored, which comprises the following steps:
Step 1, observing matrix is constructed;
Step 2, SAR image sparse information Restoration model is constructed according to the observing matrix that step 1 obtains;
Step 3, the strategy design that SAR image sparse information Restoration model obtained in step 2 is grouped according to variable is based on
The Distributed Problem Solving Algorithm of ADMM algorithm;
Step 4, the Distributed Problem Solving Algorithm based on ADMM algorithm obtained using the design in Spark platform building step 3, most
SAR image to be restored is obtained eventually.
2. a kind of distributed method that effectively SAR image based on ADMM is restored according to claim 1, feature exist
In in step 1, the construction method of observing matrix is: by two-dimentional sampled echo and two dimension image vector to be reconstructed.
3. a kind of distributed method that effectively SAR image based on ADMM is restored according to claim 1, feature exist
In, in step 2, the mathematic(al) representation of SAR image sparse information Restoration model are as follows:
S.t.y=Ax
Wherein, x ∈ RnIndicate reconstruction signal, A ∈ Rm×nFor the observing matrix of construction, y ∈ RmIndicate the signal observed;||·|
|11 norm of representing matrix.
4. a kind of distributed method that effectively SAR image based on ADMM is restored according to claim 1, feature exist
In, in step 3, to SAR image sparse information Restoration model obtained in step 2 according to variable be grouped strategy design be based on
The Distributed Problem Solving Algorithm of ADMM algorithm, specifically includes:
S1, the SAR image sparse information Restoration model according to obtained in step 2 construct the SAR image sparse information of variable grouping
Restore distributed model;
S2 restores distributed model to SAR image sparse information using distributed ADMM algorithm and solves, obtains based on ADMM
The Distributed Problem Solving Algorithm of algorithm.
5. a kind of distributed method that effectively SAR image based on ADMM is restored according to claim 4, feature exist
In in S1, the SAR image sparse information Restoration model according to obtained in step 2 constructs the SAR image sparse information of variable grouping
Restore distributed model, specifically include:
Setting SAR image sparse information Restoration model includes N number of independent subproblem, each independent subproblem is carried out parallel
Change and solve, the SAR image sparse information for obtaining variable grouping restores distributed model;SAR image sparse information restores distributed
The mathematic(al) representation of model are as follows:
Wherein, A=[A1;A2;…;AN],λ is regularization parameter.
6. a kind of distributed method that effectively SAR image based on ADMM is restored according to claim 4, feature exist
In, in S2, the expression formula of the Distributed Problem Solving Algorithm based on ADMM algorithm are as follows:
Wherein, Sλ(a) it is expressed as soft threshold operator to act on the vector that each component of a obtains, the following institute of concrete form
Show:
giIndicate gradient,τ, β, γ are initial
Change parameter, pnFor iteration m ultiple.
7. a kind of distributed method that effectively SAR image based on ADMM is restored according to claim 1, feature exist
In, in step 4, the Distributed Problem Solving Algorithm based on ADMM algorithm obtained using the design in Spark platform building step 3,
It specifically includes:
S1, initiation parameter, including regularization parameter λ, τ, beta, gamma, step-length stepsize and iteration precision ε > 0, provide simultaneously
Initial point x0,p0,res0;
S2 utilizes the distributed multiplication computation residual error res=Ax-b of matrix and vector;
S3 utilizes the distributed multiplication computation gradient g=A of matrix transposition and vector*(Ax-b-p/ β), wherein p is algorithm multiplier;
S4 is updated x=soft (x-stepsizeg) using threshold operator;
S5, if | | xk+1-xk| |≤ε, then iteration stopping, exports xk, wherein xkFor the SAR image restored;Otherwise it goes to
S2。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910544976.8A CN110244299B (en) | 2019-06-21 | 2019-06-21 | Distributed method for SAR image recovery based on ADMM |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910544976.8A CN110244299B (en) | 2019-06-21 | 2019-06-21 | Distributed method for SAR image recovery based on ADMM |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110244299A true CN110244299A (en) | 2019-09-17 |
CN110244299B CN110244299B (en) | 2021-12-28 |
Family
ID=67888947
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910544976.8A Active CN110244299B (en) | 2019-06-21 | 2019-06-21 | Distributed method for SAR image recovery based on ADMM |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110244299B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111948652A (en) * | 2020-07-17 | 2020-11-17 | 北京理工大学 | SAR intelligent parameterization super-resolution imaging method based on deep learning |
CN112272028A (en) * | 2020-09-23 | 2021-01-26 | 中国人民解放军空军预警学院 | Two-dimensional sparse signal reconstruction method based on matrix approximate message transmission |
CN113671502A (en) * | 2021-08-17 | 2021-11-19 | 桂林电子科技大学 | Distributed compressed sensing parallel imaging method of synthetic aperture radar |
JPWO2021234797A1 (en) * | 2020-05-18 | 2021-11-25 |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102854504A (en) * | 2011-06-30 | 2013-01-02 | 中国科学院电子学研究所 | Method for sparse synthetic aperture radars imaging on basis of echo simulation operators |
CN104077751A (en) * | 2014-06-19 | 2014-10-01 | 北京航空航天大学 | Hyperspectral image reconstruction method based on mixed norm tracing algorithm |
WO2015112876A1 (en) * | 2014-01-23 | 2015-07-30 | Westerngeco Llc | Large survey compressive designs |
CN105931195A (en) * | 2016-04-11 | 2016-09-07 | 华中科技大学 | Synthetic aperture radar image noise suppression method |
CN106355196A (en) * | 2016-08-23 | 2017-01-25 | 大连理工大学 | Method of identifying synthetic aperture radar image targets based on coupled dictionary learning |
US20170350974A1 (en) * | 2016-04-29 | 2017-12-07 | The Boeing Company | Methods and systems for model based automatic target recognition in sar data |
US20180198992A1 (en) * | 2016-08-30 | 2018-07-12 | Shenzhen University | Infrared image reconstruction method based on block-sparse compressive sensing and system thereof |
CN109597075A (en) * | 2018-12-29 | 2019-04-09 | 内蒙古工业大学 | A kind of imaging method and imaging device based on thinned array |
CN109884637A (en) * | 2019-04-02 | 2019-06-14 | 中国民航大学 | SAR sparse feature based on approximate message transmission enhances imaging method |
-
2019
- 2019-06-21 CN CN201910544976.8A patent/CN110244299B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102854504A (en) * | 2011-06-30 | 2013-01-02 | 中国科学院电子学研究所 | Method for sparse synthetic aperture radars imaging on basis of echo simulation operators |
WO2015112876A1 (en) * | 2014-01-23 | 2015-07-30 | Westerngeco Llc | Large survey compressive designs |
CN104077751A (en) * | 2014-06-19 | 2014-10-01 | 北京航空航天大学 | Hyperspectral image reconstruction method based on mixed norm tracing algorithm |
CN105931195A (en) * | 2016-04-11 | 2016-09-07 | 华中科技大学 | Synthetic aperture radar image noise suppression method |
US20170350974A1 (en) * | 2016-04-29 | 2017-12-07 | The Boeing Company | Methods and systems for model based automatic target recognition in sar data |
CN106355196A (en) * | 2016-08-23 | 2017-01-25 | 大连理工大学 | Method of identifying synthetic aperture radar image targets based on coupled dictionary learning |
US20180198992A1 (en) * | 2016-08-30 | 2018-07-12 | Shenzhen University | Infrared image reconstruction method based on block-sparse compressive sensing and system thereof |
CN109597075A (en) * | 2018-12-29 | 2019-04-09 | 内蒙古工业大学 | A kind of imaging method and imaging device based on thinned array |
CN109884637A (en) * | 2019-04-02 | 2019-06-14 | 中国民航大学 | SAR sparse feature based on approximate message transmission enhances imaging method |
Non-Patent Citations (5)
Title |
---|
BO ZHAO 等: ""An Admm Based Target Reconstruction Approach for Deceptively Jammed Sar"", 《2018 IEEE 10TH SENSOR ARRAY AND MULTICHANNEL SIGNAL PROCESSING WORKSHOP (SAM)》 * |
LINA SUN 等: ""Approximate Kernel Regression Based on Distributed ADMM Algorithm"", 《RECENT TRENDS IN INTELLIGENT COMPUTING, COMMUNICATION AND DEVICES. PROCEEDINGS OF ICCD 2018. ADVANCES IN INTELLIGENT SYSTEMS AND COMPUTING (1031)》 * |
SAUPTIK DHAR 等: ""ADMM based scalable machine learning on Spark"", 《2015 IEEE INTERNATIONAL CONFERENCE ON BIG DATA (BIG DATA)》 * |
杨俊刚: ""利用稀疏信息的正则化雷达成像理论与方法研究"", 《中国博士学位论文全文数据库 信息科技辑》 * |
马键 等: ""基于双重群组套索的高维空间多值处置效应估计"", 《***工程理论与实践》 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPWO2021234797A1 (en) * | 2020-05-18 | 2021-11-25 | ||
WO2021234797A1 (en) * | 2020-05-18 | 2021-11-25 | 日本電信電話株式会社 | Learning method, resolution enhancement method, learning device, and computer program |
JP7425362B2 (en) | 2020-05-18 | 2024-01-31 | 日本電信電話株式会社 | Learning method, high resolution method, learning device and computer program |
CN111948652A (en) * | 2020-07-17 | 2020-11-17 | 北京理工大学 | SAR intelligent parameterization super-resolution imaging method based on deep learning |
CN112272028A (en) * | 2020-09-23 | 2021-01-26 | 中国人民解放军空军预警学院 | Two-dimensional sparse signal reconstruction method based on matrix approximate message transmission |
CN113671502A (en) * | 2021-08-17 | 2021-11-19 | 桂林电子科技大学 | Distributed compressed sensing parallel imaging method of synthetic aperture radar |
CN113671502B (en) * | 2021-08-17 | 2023-07-11 | 桂林电子科技大学 | Distributed compressed sensing parallel imaging method of synthetic aperture radar |
Also Published As
Publication number | Publication date |
---|---|
CN110244299B (en) | 2021-12-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110244299A (en) | A kind of distributed method that the SAR image based on ADMM is restored | |
CN110501072B (en) | Reconstruction method of snapshot type spectral imaging system based on tensor low-rank constraint | |
CN105931264B (en) | A kind of sea infrared small target detection method | |
Xu et al. | High quality remote sensing image super-resolution using deep memory connected network | |
CN109584319A (en) | A kind of compression of images sensing reconstructing algorithm based on non-local low rank and full variation | |
CN106780372A (en) | A kind of weight nuclear norm magnetic resonance imaging method for reconstructing sparse based on Generalized Tree | |
CN107170019B (en) | Rapid low-storage image compression sensing method | |
CN109557540B (en) | Total variation regularization correlation imaging method based on non-negative constraint of target scattering coefficient | |
CN102708576A (en) | Method for reconstructing partitioned images by compressive sensing on the basis of structural dictionaries | |
CN111861884B (en) | Satellite cloud image super-resolution reconstruction method based on deep learning | |
CN105741333B (en) | A kind of Video-SAR image Real Time Compression reconstructing method | |
CN107154021B (en) | Image super-resolution method based on deep layer thresholding convolutional neural networks | |
Hu et al. | Inverse synthetic aperture radar imaging using a deep ADMM network | |
CN111650579B (en) | InSAR mining area three-dimensional deformation estimation method and device for rock migration parameter adaptive acquisition and medium | |
CN105488759B (en) | A kind of image super-resolution rebuilding method based on local regression model | |
CN109146787A (en) | A kind of real-time reconstruction method of the double camera spectrum imaging system based on interpolation | |
CN104199627A (en) | Gradable video coding system based on multi-scale online dictionary learning | |
CN107705342A (en) | A kind of infrared image reconstructing method based on self-adapting generalized orthogonal matching pursuit | |
CN103473744B (en) | Spatial domain based on the sampling of variable weight formula compressed sensing can downscaled images reconstructing method | |
CN104933745A (en) | Correlated imaging method based on fractal interpolation for improving image resolution | |
WO2022193379A1 (en) | Image reconstruction model generation method and apparatus, image reconstruction method and apparatus, device, and medium | |
CN111007509A (en) | Inverse synthetic aperture radar two-dimensional super-resolution imaging method | |
CN104091364A (en) | Single-image super-resolution reconstruction method | |
Ma et al. | Super-resolution geomagnetic reference map reconstruction based on dictionary learning and sparse representation | |
CN111161370B (en) | Human body multi-core DWI joint reconstruction method based on AI |
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 |