CN104657944B - Based on the compressed sensing remote sensing images method for reconstructing with reference to image texture constraint - Google Patents
Based on the compressed sensing remote sensing images method for reconstructing with reference to image texture constraint Download PDFInfo
- Publication number
- CN104657944B CN104657944B CN201410851641.8A CN201410851641A CN104657944B CN 104657944 B CN104657944 B CN 104657944B CN 201410851641 A CN201410851641 A CN 201410851641A CN 104657944 B CN104657944 B CN 104657944B
- Authority
- CN
- China
- Prior art keywords
- image
- target image
- texture
- feature vector
- constraint
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Abstract
The invention discloses a kind of compressed sensing remote sensing images method for reconstructing based on reference to image texture constraint, includes that initial value is arranged to the sparse coefficient of pre-set target image, calculate the preconfigured sparse coefficient with reference to image small echo sub-image;Based on texture in wavelet transformed domain statistical property, target image and the texture feature vector with reference to image small echo sub-image are calculated;Based on Canberra criterion distances, the distance of above-mentioned two texture feature vector is calculated, similarity is obtained;According to reference to constraint type, similarity is added in the sparse coefficient of target image and is updated, calculated texture feature vector similarity, obtain the constraint degree of target image;Renewal process is repeated, until target image sparse coefficient meets pre-set stopping criterion for iteration.Beneficial effects of the present invention are:Using the Statistic Texture with reference to image wavelet transformation sub-image as prior-constrained, it is added in the reconstruction process of target image, improves the reconstruction precision of target image.
Description
Technical field
The present invention relates to the compressed sensing signal reconstruction methods of multi- source Remote Sensing Data data, it particularly relates to a kind of based on ginseng
Examine the compressed sensing remote sensing images method for reconstructing of image texture constraint.
Background technology
In remote sensing image application, the same area generally comprises the image of multi-source, multidate, the spectrum between these images
Although different, there are prodigious similitudes for its texture.As one of remote sensing image identification feature, image texture structure and ground
Object light spectrum signature and shape feature are used for the identification of remote sensing image together.Compared with the spectral signature of Target scalar, remote sensing shadow
The texture and structural characteristic of atural object is relatively more stablized as in, therefore is of great significance in high resolution image analysis.
In consideration of it, one punishment bound term of construction is needed, it can be with the texture information with reference to image come constrained objective image
Reconstruction process;It follows that being badly in need of a kind of compressed sensing reconstruction algorithm based on reference to image texture constraint now, by borrowing
Human visual system is reflected to the processing procedure of image, calculates target image and the statistics spy with reference to image texture in wavelet coefficient
Sign, builds corresponding feature vector respectively, with the similarity degree structure of feature vector with reference to constraint, the sparse system of constrained objective image
Several reconstruction process improves the precision of reconstructed image.
Invention content
The object of the present invention is to provide it is a kind of based on reference to image texture constraint compressed sensing remote sensing images method for reconstructing,
This method uses for reference processing procedure of the human visual system to image, calculates target image and the system with reference to image texture wavelet coefficient
Feature is counted, builds corresponding feature vector respectively, with the similarity degree structure of feature vector with reference to constraint, constrained objective image is dilute
The reconstruction process of sparse coefficient improves the precision of reconstructed image.
The purpose of the present invention is be achieved through the following technical solutions:
A kind of compressed sensing remote sensing images method for reconstructing based on reference to image texture constraint, includes the following steps:
Step 1:Sparse coefficient setting initial value to pre-set target image is full null vector, and calculating is pre-configured with
The reference image small echo sub-image to match with target image sparse coefficient;
Step 2:Based on texture in the statistical property of wavelet transformed domain, according to the sparse coefficient of target image and image is referred to
The sparse coefficient of small echo sub-image calculates the target image and the texture feature vector with reference to image small echo sub-image;
Step 3:Based on Canberra criterion distances, calculates the texture feature vector of target image and refer to the small marble of image
The distance of the texture feature vector of image is to get to the similarity of above-mentioned two texture feature vector;
Step 4:According to pre-defined reference constraint type, similarity is added in the sparse coefficient of target image into
Row update calculates the similarity of the texture feature vector and the texture feature vector with reference to image of updated target image, root
Target image is obtained based on the constraint degree with reference to image according to similarity;
Step 5:Step 2 is repeated to the process of step 4, until the sparse coefficient of the target image meet it is pre-set
Stopping criterion for iteration.
Further, in step 2, the texture feature vector of the target image includes energy, standard deviation, average absolute
Deviation and entropy.
Further, in step 2, the texture feature vector of the wavelet transformation sub-image includes energy, standard deviation, puts down
Equal absolute deviation and entropy.
Further, in step 4, the texture feature vector of the texture feature vector of the target image and reference image
Similarity and target image based on the constraint degree with reference to image inversely.
Beneficial effects of the present invention are:About using the Statistic Texture with reference to image wavelet transformation sub-image as priori
Beam is added in the reconstruction process of target image, improves the reconstruction precision of target image.
Description of the drawings
It in order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, below will be to institute in embodiment
Attached drawing to be used is needed to be briefly described, it should be apparent that, the accompanying drawings in the following description is only some implementations of the present invention
Example, for those of ordinary skill in the art, without creative efforts, can also obtain according to these attached drawings
Obtain other attached drawings.
Fig. 1 is a kind of compressed sensing remote sensing images based on reference to image texture constraint according to embodiments of the present invention
The flow chart of method for reconstructing.
Specific implementation mode
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation describes, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, the every other embodiment that those of ordinary skill in the art are obtained belong to what the present invention protected
Range.
As shown in Figure 1, a kind of compressed sensing remote sensing figure based on reference to image texture constraint according to the ... of the embodiment of the present invention
As method for reconstructing, include the following steps:
Step 1:Sparse coefficient setting initial value to pre-set target image is full null vector, and calculating is pre-configured with
The reference image small echo sub-image to match with target image sparse coefficient;
Step 2:Based on texture in the statistical property of wavelet transformed domain, according to the sparse coefficient of target image and image is referred to
The sparse coefficient of small echo sub-image calculates the target image and the texture feature vector with reference to image small echo sub-image;
Step 3:Based on Canberra criterion distances, calculates the texture feature vector of target image and refer to the small marble of image
The distance of the texture feature vector of image is to get to the similarity of above-mentioned two texture feature vector;
Step 4:According to pre-defined reference constraint type, similarity is added in the sparse coefficient of target image into
Row update calculates the similarity of the texture feature vector and the texture feature vector with reference to image of updated target image, root
Target image is obtained based on the constraint degree with reference to image according to similarity;
Step 5:Step 2 is repeated to the process of step 4, until the sparse coefficient of the target image meet it is pre-set
Stopping criterion for iteration.
Wherein, in step 2, the texture feature vector of the target image includes energy, standard deviation, mean absolute deviation
And entropy.
In step 2, the texture feature vector of the wavelet transformation sub-image includes that energy, standard deviation, average absolute are inclined
Difference and entropy.
In addition, in step 4, the phase of the texture feature vector of the target image and the texture feature vector with reference to image
Inversely like the constraint degree of degree and target image based on reference image, i.e., similarity is higher, and constraint degree is smaller;Similarity is got over
Low, constraint degree is bigger.
When specifically used,
Step 1:If the initial value of the sparse coefficient of target image is full null vectorα 0 obj , calculate the sparse system with reference to image
Numberα ref ;
To make full use of the perception characteristics of human visual system, we are become when carrying out sparse representation to image using small echo
Change, multiresolution, multi-angle time-frequency characteristic so that our reason feature extracting method can be multiple dimensioned, multidirectional
Textural characteristics.
Step 2:Statistical nature based on texture in wavelet coefficient calculates target image and with reference to image wavelet coefficient
Statistical nature and feature vector fv obj , fv ref ;
We assume that the Energy distribution of scale space can be used as unique texture description feature, the psychology of visual cortex
This judgement, therefore the energy of wavelet transformation sub-image are also supported in researchF energy , standard deviationF sd , mean absolute deviationF aad And
EntropyF entropy,These parameters can be used as textural characteristics identification parameter,
;
;
;
;
Whereinα(x,y)For the sub-image after wavelet transformation,
,
Indicate the mean value of sub-image.
It follows that the target image and the texture feature vector with reference to image small echo sub-image are respectively:
;
;
Step 3:Based on Canberra distances, calculate the distance i.e. similarity degree W of feature vector, wherein Canberra away from
From being defined as follows:
Wherein, m is the number of small echo sub-image.
Step 4:Definition adds it to target image sparse coefficient with reference to the form constrainedα obj Renewal process, make
When proper target image and reference image close proximity, constraint is smaller, conversely, constraint is larger;
It is as follows according to pre-defined reference constraint:
Wherein, matrix IiIndicate i-th of small echo sub-image of extraction,W i For target image and refer to i-th of sub-image line of image
Manage the distance between feature.It is entire to be with reference to the meaning constrained:If target image and reference i-th of small echo sub-image similarity of image
It is relatively low, then the distance between two width image texture featuresW i It is larger, it should to retain otherness.
Object function becomes as a result,:
Step 5:The above object function is solved, target image sparse coefficient is acquired, repeats the process of step 2- steps 4, directly
Sparse coefficient to target image is met the requirements.
In conclusion by means of the above-mentioned technical proposal of the present invention, processing procedure of the human visual system to image is used for reference,
Target image and the statistical nature with reference to image texture in wavelet coefficient are calculated, corresponding feature vector is built respectively, with spy
For the similarity degree structure of sign vector with reference to constraint, the reconstruction process of constrained objective image sparse coefficient improves the essence of reconstructed image
Degree.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
With within principle, any modification, equivalent replacement, improvement and so on should all be included in the protection scope of the present invention god.
Claims (4)
1. a kind of compressed sensing remote sensing images method for reconstructing based on reference to image texture constraint, which is characterized in that including following
Step:
Step 1:Be full null vector to the sparse coefficient of pre-set target image setting initial value, calculate it is preconfigured with
The sparse coefficient for the reference image small echo sub-image that target image matches;
Step 2:Based on texture in the statistical property of wavelet transformed domain, according to the sparse coefficient of target image and image small echo is referred to
The sparse coefficient of sub-image calculates the target image and the texture feature vector with reference to image small echo sub-image;
Step 3:Based on Canberra criterion distances, calculates the texture feature vector of target image and refer to image small echo sub-image
Texture feature vector distance to get to the similarity of above-mentioned two texture feature vector;
Step 4:According to pre-defined reference constraint type, similarity is added in the sparse coefficient of target image and is carried out more
Newly, the texture feature vector for calculating updated target image is similar to the texture feature vector with reference to image small echo sub-image
Degree obtains target image based on the constraint degree with reference to image according to similarity;
Step 5:Step 2 is repeated to the process of step 4, until the sparse coefficient of the target image meets pre-set iteration
End condition.
2. the compressed sensing remote sensing images method for reconstructing according to claim 1 based on reference to image texture constraint, special
Sign is, in step 2, the texture feature vector of the target image include energy, standard deviation, mean absolute deviation and
Entropy.
3. the compressed sensing remote sensing images method for reconstructing according to claim 1 based on reference to image texture constraint, special
Sign is that in step 2, the texture feature vector with reference to image small echo sub-image includes energy, standard deviation, average absolute
Deviation and entropy.
4. the compressed sensing remote sensing images method for reconstructing according to claim 1 based on reference to image texture constraint, special
Sign is, in step 4, the texture feature vector of the target image and the texture feature vector with reference to image small echo sub-image
Similarity and target image based on the constraint degree with reference to image inversely.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410851641.8A CN104657944B (en) | 2014-12-31 | 2014-12-31 | Based on the compressed sensing remote sensing images method for reconstructing with reference to image texture constraint |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410851641.8A CN104657944B (en) | 2014-12-31 | 2014-12-31 | Based on the compressed sensing remote sensing images method for reconstructing with reference to image texture constraint |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104657944A CN104657944A (en) | 2015-05-27 |
CN104657944B true CN104657944B (en) | 2018-08-14 |
Family
ID=53249022
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410851641.8A Active CN104657944B (en) | 2014-12-31 | 2014-12-31 | Based on the compressed sensing remote sensing images method for reconstructing with reference to image texture constraint |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104657944B (en) |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105551000B (en) * | 2015-12-11 | 2019-09-20 | 中国科学院遥感与数字地球研究所 | Remote sensing images method for reconstructing based on the constraint of reference image and the constraint of non-convex low-rank |
CN106023274B (en) * | 2016-01-16 | 2019-03-15 | 中国科学院遥感与数字地球研究所 | A kind of compressed sensing image rebuilding method of combination expert filter sparse constraint |
CN105741240A (en) * | 2016-01-18 | 2016-07-06 | 中国科学院遥感与数字地球研究所 | Remote sensing image reconstruction method based on reference image texture constraint and non-convex low-rank constraint |
CN107784278B (en) * | 2017-10-18 | 2019-02-05 | 清华大学 | Sparse image reconstruction accuracy reduction complexity method is improved with structuring is prior-constrained |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103455988A (en) * | 2013-09-17 | 2013-12-18 | 西安电子科技大学 | Super-resolution image reconstruction method based on structure self-similarity and sparse representation |
CN103824263A (en) * | 2014-03-03 | 2014-05-28 | 哈尔滨工业大学 | Remote sensing image sparseness estimation method based on mixing transformation |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8570442B2 (en) * | 2011-07-12 | 2013-10-29 | Xerox Corporation | Hyperspectral image reconstruction via a compressed sensing framework |
-
2014
- 2014-12-31 CN CN201410851641.8A patent/CN104657944B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103455988A (en) * | 2013-09-17 | 2013-12-18 | 西安电子科技大学 | Super-resolution image reconstruction method based on structure self-similarity and sparse representation |
CN103824263A (en) * | 2014-03-03 | 2014-05-28 | 哈尔滨工业大学 | Remote sensing image sparseness estimation method based on mixing transformation |
Non-Patent Citations (5)
Title |
---|
COMPRESSED SENSING BASED REMOTE SENSING IMAGE RECONSTRUCTION USING AN AUXILIARY IMAGE AS PRIORS;Hao Geng et al;《IEEE Geoscience and Remote Sensing Symposium》;20141106;全文 * |
COMPRESSIVE IMAGING OF COLOR IMAGES;Pradeep Nagesh et al;《2009 IEEE International Conference on Acoustics, Speech and Signal Processing》;20091231;全文 * |
Remote Sensing Image Fusion via Sparse Representations Over Learned Dictionaries;Shutao Li et al;《IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING》;20130930;第51卷(第9期);全文 * |
Texture classification with combined rotation and scale invariant wavelet features;K.Muneeswaran et al;《Pattern Recognition》;20051231(第38期);全文 * |
一种多源多光谱遥感图像特征点匹配技术;刘志文 等;《科技导报》;20131231;第31卷(第15期);全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN104657944A (en) | 2015-05-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104657944B (en) | Based on the compressed sensing remote sensing images method for reconstructing with reference to image texture constraint | |
Liu et al. | Simultaneous image fusion and denoising with adaptive sparse representation | |
Razdan et al. | A hybrid approach to feature segmentation of triangle meshes | |
Alvarez-Ramirez et al. | Performance of a high-dimensional R/S method for Hurst exponent estimation | |
CN107689079B (en) | Cloud scene reconstruction method combining satellite cloud picture and natural image | |
WO2014021996A1 (en) | System and method for reduced incremental spectral clustering | |
CN105551000B (en) | Remote sensing images method for reconstructing based on the constraint of reference image and the constraint of non-convex low-rank | |
CN105654023B (en) | The method and apparatus for identifying object risk | |
CN111008975B (en) | Mixed pixel unmixing method and system for space artificial target linear model | |
Deng et al. | Infrared small target detection based on the self-information map | |
CN110349117A (en) | A kind of infrared image and visible light image fusion method, device and storage medium | |
Zhang et al. | A method for organs classification and fruit counting on pomegranate trees based on multi-features fusion and support vector machine by 3D point cloud | |
CN104268561B (en) | High spectrum image solution mixing method based on structure priori low-rank representation | |
CN105825217A (en) | Hyperspectral image interested area automatic extraction method based on active contour model | |
CN113446998B (en) | Hyperspectral target detection data-based dynamic unmixing method | |
CN109300115B (en) | Object-oriented multispectral high-resolution remote sensing image change detection method | |
CN105741240A (en) | Remote sensing image reconstruction method based on reference image texture constraint and non-convex low-rank constraint | |
CN111310772B (en) | Point line characteristic selection method and system for binocular vision SLAM | |
CN116755005A (en) | Model-based method and system for monitoring magnetic field during lightning generation | |
CN108304766B (en) | A method of dangerous material stockyard is screened based on high-definition remote sensing | |
CN111062888A (en) | Hyperspectral image denoising method based on multi-target low-rank sparsity and spatial-spectral total variation | |
CN107607927B (en) | Stripe laser echo information extraction method | |
Papadomanolaki et al. | Patch-based deep learning architectures for sparse annotated very high resolution datasets | |
CN108280466B (en) | Polarization SAR (synthetic aperture radar) feature classification method based on weighted nuclear norm minimization | |
CA2873918A1 (en) | Method and system for the three-dimensional reconstruction of structures |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |