CN108460402A - Polarimetric SAR image supervised classification method and device - Google Patents
Polarimetric SAR image supervised classification method and device Download PDFInfo
- Publication number
- CN108460402A CN108460402A CN201810036297.5A CN201810036297A CN108460402A CN 108460402 A CN108460402 A CN 108460402A CN 201810036297 A CN201810036297 A CN 201810036297A CN 108460402 A CN108460402 A CN 108460402A
- Authority
- CN
- China
- Prior art keywords
- feature vector
- sar image
- polarimetric sar
- acquisition module
- vector collection
- 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
- 238000000034 method Methods 0.000 title claims abstract description 97
- 239000013598 vector Substances 0.000 claims abstract description 132
- 230000000007 visual effect Effects 0.000 claims abstract description 88
- 230000010287 polarization Effects 0.000 claims abstract description 82
- 239000011159 matrix material Substances 0.000 claims abstract description 46
- 238000013507 mapping Methods 0.000 claims abstract description 28
- 238000012549 training Methods 0.000 claims abstract description 20
- 230000009467 reduction Effects 0.000 claims abstract description 19
- 238000012545 processing Methods 0.000 claims abstract description 7
- 238000003860 storage Methods 0.000 claims description 12
- 238000010219 correlation analysis Methods 0.000 claims description 10
- 238000000354 decomposition reaction Methods 0.000 claims description 10
- 239000000126 substance Substances 0.000 claims description 9
- 101001056160 Homo sapiens Methylcrotonoyl-CoA carboxylase subunit alpha, mitochondrial Proteins 0.000 claims description 8
- 102100026552 Methylcrotonoyl-CoA carboxylase subunit alpha, mitochondrial Human genes 0.000 claims description 8
- 230000008859 change Effects 0.000 claims description 7
- 238000001914 filtration Methods 0.000 claims description 7
- 239000003086 colorant Substances 0.000 claims description 6
- 230000003044 adaptive effect Effects 0.000 claims description 4
- 230000017105 transposition Effects 0.000 claims description 3
- 239000000284 extract Substances 0.000 abstract description 7
- 238000002474 experimental method Methods 0.000 description 16
- 238000004458 analytical method Methods 0.000 description 8
- 230000000694 effects Effects 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 238000004088 simulation Methods 0.000 description 4
- 101100025318 Danio rerio mvd gene Proteins 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 240000004658 Medicago sativa Species 0.000 description 2
- 235000010624 Medicago sativa Nutrition 0.000 description 2
- 244000046052 Phaseolus vulgaris Species 0.000 description 2
- 235000010627 Phaseolus vulgaris Nutrition 0.000 description 2
- 244000061456 Solanum tuberosum Species 0.000 description 2
- 235000002595 Solanum tuberosum Nutrition 0.000 description 2
- 241000209140 Triticum Species 0.000 description 2
- 235000021307 Triticum Nutrition 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000004615 ingredient Substances 0.000 description 2
- 230000001788 irregular Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- 235000016068 Berberis vulgaris Nutrition 0.000 description 1
- 241000335053 Beta vulgaris Species 0.000 description 1
- 241000219310 Beta vulgaris subsp. vulgaris Species 0.000 description 1
- 240000002791 Brassica napus Species 0.000 description 1
- 235000004977 Brassica sinapistrum Nutrition 0.000 description 1
- 244000025254 Cannabis sativa Species 0.000 description 1
- 235000021536 Sugar beet Nutrition 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000001035 drying Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 235000012015 potatoes Nutrition 0.000 description 1
- 238000000513 principal component analysis Methods 0.000 description 1
- VMXUWOKSQNHOCA-UKTHLTGXSA-N ranitidine Chemical compound [O-][N+](=O)\C=C(/NC)NCCSCC1=CC=C(CN(C)C)O1 VMXUWOKSQNHOCA-UKTHLTGXSA-N 0.000 description 1
- 239000002689 soil Substances 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Abstract
The present invention relates to technical field of image processing, a kind of polarimetric SAR image supervised classification method and device are specifically provided, it is intended to the technical issues of solving that the various visual angles feature of polarization SAR how to be made full use of to improve the nicety of grading of image.For this purpose, the polarimetric SAR image supervised classification method in the present invention, includes the following steps：According to polarization characteristic vector set, texture feature vector collection and the color feature vector collection of training sample, its corresponding dimensionality reduction mapping matrix is obtained using default various visual angles sub-space learning model；According to acquired each dimensionality reduction mapping matrix, the low-dimensional feature of each pixel is obtained；According to the low-dimensional feature of acquired each pixel, classified to pixel to be sorted using default grader.The various visual angles feature that polarization SAR can be made full use of through the invention extracts feature while the structure for keeping data, discriminant information and Viewing-angle information, improves the nicety of grading of image.
Description
Technical field
The present invention relates to technical field of image processing, and in particular to a kind of polarimetric SAR image supervised classification method and dress
It sets.
Background technology
Polarimetric synthetic aperture radar (Synthetic Aperture Radar, SAR) is combined using different POLARIZATION CHANNELs,
More more rich than traditional SAR, more detailed terrestrial object information can be obtained, thus is widely used in the fields such as agricultural, military affairs, ocean.
Classification of Polarimetric SAR Image is the important content of polarization SAR image procossing and polarimetric SAR image interpretation.
The feature that existing Classification of Polarimetric SAR Image method is used is mainly from polarization SAR data, such as polarization scattering matrix
S, it is extracted in coherence matrix T and polarization covariance matrix C etc. and the various Polarization target decomposition methods of polarizing.In recent years, line
Reason feature and color characteristic are also used for the feature of Classification of Polarimetric SAR Image.But due to the redundancy between big measure feature, one
A little single-view sub-space learning methods, such as principal component analysis PCA, laplacian eigenmaps LE, for extracting polarization SAR data
Feature when, have ignored the otherness and relevance between different visual angles feature.
Although various visual angles sub-space learning method avoids above-mentioned technical problem, but various visual angles sub-space learning method exists
It is difficult to keep data structure, discriminant information and Viewing-angle information simultaneously in terms of Classification of Polarimetric SAR Image, and then leads to polarization SAR figure
The nicety of grading of picture is bad.Various visual angles sub-space learning method includes mainly canonical correlation analysis CCA methods, the typical phase of various visual angles
Close analysis MCCA methods and various visual angles discriminant analysis MVDA methods.Wherein, canonical correlation analysis CCA methods can only handle two
The data at a visual angle, and this method does not utilize discriminant information.Various visual angles canonical correlation analysis MCCA methods are canonical correlations
The various visual angles for analyzing CCA methods are expanded, and this method can handle various visual angles data, but this method utilize discriminant information and
The holding of data structure.Although various visual angles discriminant analysis MVDA methods can handle various visual angles data and this process employs differentiations
Information, but this method does not account for holding data structure.
Invention content
In order to solve the above problem in the prior art, in order to solve how to make full use of the various visual angles of polarization SAR special
Sign improves the technical issues of nicety of grading of image, and the present invention provides a kind of polarimetric SAR image supervised classification method and dresses
It sets.
In a first aspect, polarimetric SAR image supervised classification method in the present invention, including：
According to polarization characteristic vector set, texture feature vector collection and the color feature vector collection for presetting training sample, and
Default truly substance markers image obtains its corresponding dimensionality reduction using default various visual angles sub-space learning model and maps square
Battle array；
According to acquired each dimensionality reduction mapping matrix, the low-dimensional feature of each pixel is obtained；
The low-dimensional feature of each pixel according to the acquisition carries out each pixel to be sorted using default grader
Classification；
Wherein, the default various visual angles sub-space learning model is according to various visual angles canonical correlation analysis MCCA method moulds
Type, and combine the holding of data structure and discriminant information and the model of adaptive weighting parameter structure：
Wherein,I=1,2 ..., v, and It is described for the sample data set at v visual angleForvThe corresponding map vector in a visual angle, it is describedFor the corresponding visual angle weight at v visual angle, the DiIt is
The dimension of i visual angle characteristic, the n are number of samples, the wiIndicate mapping matrix WiAny one column vector, andThe d is the dimension to lower, and the T indicates transposition；The r
It is parameter with β.
Preferably, the default training sample is the classification according to default truly substance markers, from polarization SAR to be sorted
The pixel of all categories of preset quantity is randomly selected in image.
Preferably, the method " according to the low-dimensional feature of each pixel, is clicking through pixel to be sorted using grader
Further include following step after the step of row classification "：
Every a kind of sample in polarimetric SAR image in classification results is identified with different colours, obtains Classification of Polarimetric SAR Image
As a result coloured picture.
Preferably, the method is in " polarization characteristic vector set, texture feature vector collection and the color spy according to training sample
Levy vector set, obtain its corresponding dimensionality reduction mapping matrix using various visual angles sub-space learning model " the step of before also wrap
Include following step：
The polarimetric SAR image to be sorted is filtered using filter；
It is obtained using polarization SAR initial data and goal decomposition method according to filtered polarimetric SAR image to be sorted
Take polarization characteristic vector set；
It is obtained using gray co-occurrence matrix and Gabor filtering methods according to filtered polarimetric SAR image to be sorted
Texture feature vector collection；
It is special to be obtained using RGB and hsv color histogram method for color according to filtered polarimetric SAR image to be sorted
Levy vector set；
The training is obtained according to acquired polarization characteristic vector set, texture feature vector collection and color feature vector collection
Polarization characteristic vector set, texture feature vector collection and the color feature vector collection of sample.
Preferably, the method further includes according to the default various visual angles sub-space learning model, and iterative solution obtains v
The corresponding mapping matrix in visual angleWith the corresponding visual angle weight at v visual angleThe specific steps are：
Described in fixationUsing Lagrange multiplier algorithm, the corresponding visual angle weight at v visual angle is obtained
Described in fixationUsing Lagrange multiplier algorithm, the corresponding d map vector w in each visual angle is obtainedi1,
wi2,…,wid, and then constitute mapping matrix
Wi=[wi1,wi2,…,wid]
Wherein, i=1 ..., v.
Polarimetric SAR image Supervised classification device in second aspect, the present invention, including：First acquisition module, second
Acquisition module and sort module；
First acquisition module is configured to according to polarization characteristic vector set, the texture feature vector for presetting training sample
It is each to obtain it using default various visual angles sub-space learning model for collection and color feature vector collection, and truly substance markers image
Self-corresponding dimensionality reduction mapping matrix；
Second acquisition module is configured to, according to each dimensionality reduction mapping matrix acquired in first acquisition module, obtain
Take the low-dimensional feature of each pixel；
The sort module is configured to the low-dimensional feature according to each pixel acquired in second acquisition module, profit
Classified to each pixel to be sorted with default grader；
Wherein, the default various visual angles sub-space learning model is shown below：
Preferably, described device further includes mark module；The mark module is configured to different colours to the classification
Being identified per a kind of sample in the classification results acquired in module, obtains Classification of Polarimetric SAR Image result coloured picture.
Preferably, described device further includes that filter module, polarization characteristic vector set acquisition module, texture feature vector collection obtain
Modulus block, color feature vector collection acquisition module and third acquisition module；
The filter module is configured to be filtered polarimetric SAR image to be sorted using filter；
The polarization characteristic vector set acquisition module is configured to according to the filtered polarization acquired in the filter module
SAR image obtains polarization characteristic vector set using polarization SAR initial data and goal decomposition method；
The texture feature vector collection acquisition module is configured to according to the filtered polarization acquired in the filter module
SAR image obtains texture feature vector collection using gray co-occurrence matrix and Gabor filtering methods；
The color feature vector collection acquisition module is configured to according to the filtered polarization acquired in the filter module
SAR image obtains color feature vector collection using RGB and hsv color histogram method；
The third acquisition module is configured to according to the polarization characteristic acquired in the polarization characteristic vector set acquisition module
Vector set, the texture feature vector collection acquired in the texture feature vector collection acquisition module and the color feature vector collection
Color feature vector collection acquired in acquisition module obtains polarization characteristic vector set, the texture feature vector of the training sample
Collection and color feature vector collection.
Storage device in the third aspect, the present invention is suitable for by processor load simultaneously wherein being stored with a plurality of program
It executes to realize the polarimetric SAR image supervised classification method described in above-mentioned technical proposal.
Processing unit in fourth aspect, the present invention, including
Processor is adapted for carrying out each program；And
Storage device is suitable for storing a plurality of program；
Described program is suitable for being loaded by processor and being executed to realize that the polarimetric SAR image described in above-mentioned technical proposal has prison
Superintend and direct sorting technique.
Compared with the immediate prior art, above-mentioned technical proposal at least has the advantages that：
1. in the polarimetric SAR image supervised classification method of the present invention, the various visual angles of polarimetric SAR image are adequately utilized
Feature overcomes such as polarization characteristic, textural characteristics and color characteristic and only utilizes figure caused by single visual angle feature in the prior art
As the inaccurate problem of classification.
2. on the basis of making full use of polarimetric SAR image various visual angles feature, data structure is also maintained well, is differentiated
Information and Viewing-angle information, information is kept not during overcoming various visual angles sub-space learning method extraction feature in the prior art
The low problem of image classification accuracy caused by comprehensively.
Description of the drawings
Fig. 1 is the key step schematic diagram of the polarimetric SAR image supervised classification method of the embodiment of the present invention；
Fig. 2 (a) is the polarization SAR pseudo color image used in emulation experiment of the embodiment of the present invention；
Fig. 2 (b) is the true category label figure of polarization SAR pseudo color image in emulation experiment of the embodiment of the present invention；
Fig. 2 (c) is the simulation result diagram that various visual angles Canonical Correlation Analysis is used in emulation experiment of the embodiment of the present invention；
Fig. 2 (d) is the simulation result diagram that various visual angles discriminant analysis method is used in emulation experiment of the embodiment of the present invention；
Fig. 2 (e) is in emulation experiment of the embodiment of the present invention using the emulation effect for having supervision Wishart grader SWC methods
Fruit is schemed；
Fig. 2 (f) is in emulation experiment of the embodiment of the present invention using the polarimetric SAR image Supervised classification side in the present invention
The simulated effect figure of method.
Specific implementation mode
The preferred embodiment of the present invention described with reference to the accompanying drawings.It will be apparent to a skilled person that this
A little embodiments are used only for explaining the technical principle of the present invention, it is not intended that limit the scope of the invention.
Below in conjunction with the accompanying drawings, polarimetric SAR image supervised classification method in the embodiment of the present invention is illustrated.
Refering to attached drawing 1, Fig. 1 illustratively shows the key step of polarimetric SAR image supervised classification method.Such as Fig. 1
Shown, polarimetric SAR image supervised classification method may include step S1, step S2 and step S3 in the present embodiment.
Step S1, according to three visual angle characteristic vector sets for presetting training sample, i.e. polarization characteristic vector set, textural characteristics
Vector set and color feature vector collection, and default truly substance markers image, utilize default various visual angles sub-space learning model
Obtain the corresponding dimensionality reduction mapping matrix of three visual angle characteristic vector sets.
Default various visual angles sub-space learning model is foundation various visual angles canonical correlation analysis MCCA method models, and is combined
The model of the holding and adaptive weighting parameter structure of data structure and discriminant information：Such as following formula (1) depicted：
Wherein,I=1,2 ..., v, and For the sample data set at v visual angle,For v visual angle
Corresponding map vector,For the corresponding visual angle weight at v visual angle, DiFor the dimension of i-th of visual angle characteristic, n is sample
This number, wiIndicate mapping matrix WiAny one column vector, andd
For the dimension to lower, T indicates transposition；R and β is parameter.
Specifically, various visual angles canonical correlation analysis MCCA method models enter shown in following formula (2) in the present embodiment：
Wherein,I=1,2 ..., v, and
Specifically, in the present embodimentFor keeping data structure and discriminant information, and visual angle
Weight α1,α2,…,αvIt is unknown, meets Condition of Non-Negative Constrains, can be adjusted according to data adaptive.Specifically, such as following formula (3)
It is shown：
Wherein, L=D-G, D are diagonal matrix, and diagonal entry isN is number of samples,
P, q is sample, GpqIt is the element of matrix G, matrix G characterizes partial structurtes and the discriminant information of data, element make
As shown in following formula (4)：
Wherein, sample p and sample q belong to k adjacent to sample, and belong to same class, CpAnd CqFor the covariance matrix of m × m,
T is parameter, distance dSRWIt is the symmetrical version for correcting Wishart distances, dSRWAs shown in following formula (5)：
Further, in this embodiment can iteratively solve according to various visual angles sub-space learning model and obtain v visual angle pair
The map vector answeredWith the corresponding visual angle weight at v visual angleSpecific steps include step S11 and step
S12。
Step S11, it is fixedThe corresponding visual angle weight at v visual angle is obtained using Lagrange multiplier algorithmAs shown in following formula (6)：
Specifically, pass through fixation in the present embodimentIt solvesIt can obtain the first optimization task such as following formula
(7) shown in：
Wherein,
Step S12, it is fixedUsing Lagrange multiplier algorithm, obtain each visual angle it is corresponding d map to
Measure wi1,wi2,…,wid, and then mapping matrix is constituted, as shown in following formula (8)：
Wi=[wi1,wi2,…,wid](8)
Wherein, i=1 ..., v.
Specifically, in the present embodiment, pass through fixationIt solvesIt can obtain the second optimization task, such as following formula
(9) shown in：
Wherein,
Using Lagrange multiplier algorithm, generalized eigenvalue decomposition task is obtained, as shown in following formula (10)：
Aw=μ B (10)
Wherein, shown in A such as following formulas (11)：
Shown in B such as following formulas (12)：
Shown in w such as following formulas (13)：
By generalized eigenvalue decomposition, the obtained corresponding feature vector of maximum d characteristic values, and then constitute v mapping
Matrix Wi=[wi1,wi2,…,wid], i=1 ..., v.
After the structure for completing default various visual angles sub-space learning model, it is also necessary to choose training sample as model
Input, obtains the corresponding dimensionality reduction mapping matrix of each set of eigenvectors.
Further, in this embodiment default training sample is the classification according to default truly substance markers, to be sorted
Polarimetric SAR image in randomly select the pixel of all categories of preset quantity.
Specifically, preset quantity can be 100 in the present embodiment, then default training sample acquired in the present embodiment
Polarization characteristic vector set, texture feature vector collection and color feature vector collection are respectively X1、X2And X3, and X1∈R55×100,X2∈
R56×100,X3∈R40×100, the mapping matrix obtained corresponding to three set of eigenvectors is respectively W1、W2And W3。
Further, in this embodiment obtaining three feature vectors using default various visual angles sub-space learning model
It further includes before step A1, step A2, step A3, step A4 and step A5 to collect corresponding dimensionality reduction mapping matrix.
Step A1 is filtered polarimetric SAR image to be sorted using filter.
Specifically, the exquisite Lee filters that filter window size is 7*7 pixels may be used in the present embodiment, treat point
All pixels point in the polarimetric SAR image of class is filtered, and obtains filtered polarimetric SAR image.
Step A2, according to filtered polarimetric SAR image to be sorted, using polarization SAR initial data and goal decomposition
Method obtains polarization characteristic vector set.
Specifically, the step of polarization characteristic vector set is obtained in the present embodiment specifically includes：Step A21, step A22, step
Rapid A23 and step A24.
Step A21 extracts in the value of real part and imaginary values and diagonal line of 3 elements of upper triangle of covariance matrix C 3
Element constitutes 9 polarization characteristics；And extract the value of real part and imaginary values and diagonal line of 3 elements of upper triangle of coherence matrix T
Upper 3 elements constitute 9 polarization characteristics.
Step A22,6 polarization characteristics that simple transformation of the extraction based on initial data obtains, including the channels VV and HH, HV
With the channels HH, the ratio of the backscattering coefficient in the channels HV and VV, the phase difference in the channels HH-VV depolarizes ratio, total work
Rate.
Step A23,3 polarization characteristics that extraction Pauli is decomposed：Odd times scattered power | a |2, even scattered power | b |2,
Even scattered power | c |2π/4；
Extract 3 polarization characteristics that Krogager is decomposed：Sphere scattered power | ks|2, dihedron scattered power | kd|2With
Spiral volume scattering power | kh|2；
Extract 6 polarization characteristics that Freeman is decomposed：The corresponding coefficient f of surface scattering items, even scattering item it is corresponding
Coefficient fd, the corresponding coefficient f of volume scattering itemv, area scattering power Ps, even scattered power PdWith volume scattering power Pv；
Extract 6 polarization characteristics that H/A/alpha is decomposed：Eigenvalues Decomposition is carried out to covariance matrix or coherence matrix to obtain
Three eigenvalue λs arrived1,λ2,λ3And the H entropys being calculated based on three characteristic values, anisotropy coefficient A and average scattering
Angle
Extract 10 polarization characteristics that Huynen is decomposed：Goal rule A0, smooth and projection portion total volume scattering power,
The irregular B of target0, coarse and non-projection portion total scattering power, the total scattered power B of target asymmetry ingredient0- B, target
The total scattered power B of irregular ingredient0+ B, target linear degree C, the curvature D of target, the torsion resistance E of target, the spiral of target
Property F, the overlapping degree G of target, the direction H of target；Wherein, three eigenvalue λs that Van Zyl are decomposed1,λ2,λ3As 3 poles
Change feature.
The polarization characteristic of each pixel is expressed as the vector of 55*1 by step A24.
Step A3, according to filtered polarimetric SAR image to be sorted, using gray co-occurrence matrix and the filtering sides Gabor
Method obtains texture feature vector collection.
Specifically, the step of texture feature vector collection is obtained in the present embodiment specifically includes：Step A31, step A32 and step
Rapid A33.
Step A31 calculates the gray co-occurrence matrix that distance is 1 pixel and 4 directions (0 °, 45 °, 90 °, 135 °),
And then 4 energy, entropy, correlation and contrast features are obtained, amount to 16 features.
Step A32 calculates 40 picture amplitude mean values under 5 sizes and 8 directions by Gabor filtering methods, makees
It is characterized.
The textural characteristics of each pixel are expressed as the vector of 56*1 by step A33.
Step A4 is obtained according to filtered polarimetric SAR image to be sorted using RGB and hsv color histogram method
Take color feature vector collection.
Specifically, the step of color feature vector collection is obtained in the present embodiment specifically includes：Step A41, step A42 and step
Rapid A43.
Step A41, the grey level histogram in three channels based on RGB and hsv color space, calculate mean value, variance, partially
6 degree, kurtosis, energy and entropy features, in total 36 features.
Step A42 calculates 4 dominant color ratios in hsv color space, as feature.
The color characteristic of each pixel is expressed as the vector of 40*1 by step A43.
Step A5 is obtained according to acquired polarization characteristic vector set, texture feature vector collection and color feature vector collection
Polarization characteristic vector set, texture feature vector collection and the color feature vector collection of training sample.
Step S2 obtains the low-dimensional feature of each pixel according to acquired each dimensionality reduction mapping matrix.
Specifically, polarization characteristic vector set, the texture feature vector collection of pixel to be sorted acquired in the present embodiment
It is respectively Y with color feature vector collection1、Y2And Y3, the polarization characteristic of default training sample acquired in step S1 can be utilized
Vector set, texture feature vector collection and the corresponding dimensionality reduction mapping matrix W of color feature vector collection1、W2And W3, and pass throughIt calculates, obtains the low-dimensional feature of each pixel：
Step S3, according to the low-dimensional feature of each pixel obtained, using default grader to each pixel to be sorted
Classify.
Specifically, the low-dimensional feature based on each pixel acquired in step 2 in the present embodiment, utilizes default grader
Classify to pixel to be sorted.Wherein, default grader can be support vector machines grader.
Can also include step S4 further, in this embodiment after classifying to pixel to be sorted:
Step S4 identifies every a kind of sample in polarimetric SAR image in classification results with different colours, obtains polarization SAR
Image classification result coloured picture.
The effect of the present invention can be described in detail by following emulation experiments.
The present invention emulation experiment be Intel Xeib 2.6Ghz CPU, 64G Ram server hardware environment and
It is carried out under the software environment of MATLAB R2015a.
Fig. 2 (a) is the polarization SAR pseudo color image used in the emulation experiment of the embodiment of the present invention, and Fig. 2 (b) is emulation
The true category label figure of polarization SAR pseudo color image in experiment, Fig. 2 (c) are in emulation experiment using various visual angles canonical correlation point
The simulation result diagram of analysis method MCCA, Fig. 2 (d) are the simulation result diagrams that various visual angles discriminant analysis method is used in emulation experiment,
Fig. 2 (e) is using the simulated effect figure for having supervision Wishart graders SWC in emulation experiment, and Fig. 2 (f) is in emulation experiment
Using the simulated effect figure of the polarimetric SAR image supervised classification method in the present invention.As shown in Fig. 2 (a), which is the U.S.
The data in the areas Dutch Flevoland that the AIRSAR systems of Space Agency jet propulsion laboratory (NASA/JPL) obtain, position
It is one four full polarimetric SAR data regarded in L-band, original size 750*1024 is its 200*320 for emulation experiment
Subgraph, have been subjected to and be filtered, the region include 9 class atural objects：Stem beans (Stem Beans), potato (Potatoes), lucerne
Mu (Lucerne), winter wheat 1 (Winter Wheat I), winter wheat 2 (Winter Wheat II), bare area (Bare
Soil), beet (Sugar Beet), rape (Rapeseed), meadow (Grass).
Polarimetric SAR image to be sorted is divided into 9 classes by the emulation experiment of the present invention.
Fig. 2 (c), Fig. 2 (d), Fig. 2 (e) and Fig. 2 (f) are compared respectively, it can be seen that using the polarization SAR in the present invention
Image supervised classification method, compared to MCCA, MvDA and SWC, wrong point of miscellaneous point is less in region, and region consistency is preferable.
Various visual angles Canonical Correlation Analysis, various visual angles discriminant analysis method have supervision Wishart grader SWC methods
Classification accuracy rate is counted with polarimetric SAR image supervised classification method in the present invention, the results are shown in Table 1：
Table 1
| Emulation mode | Classification accuracy |
| Various visual angles Canonical Correlation Analysis | 0.9067 |
| Various visual angles discriminant analysis method | 0.9318 |
| There are supervision Wishart grader SWC methods | 0.9268 |
| The present invention | 0.9544 |
From table 1 it follows that be improved on classification accuracy rate compared to other three kinds of methods with the method for the present invention,
This is primarily due to the present invention and takes full advantage of various visual angles feature and the relationship between them, to improve image classification
Nicety of grading.
Based on technical concept identical with polarimetric SAR image supervised classification method embodiment, the embodiment of the present invention also carries
A kind of polarimetric SAR image Supervised classification device is supplied.The polarimetric SAR image Supervised classification device is carried out specifically below
It is bright.
Polarimetric SAR image Supervised classification device can also include the first acquisition module, the second acquisition module in this implementation
And sort module.
Wherein, the first acquisition module is configurable to according to polarization characteristic vector set, the textural characteristics for presetting training sample
Vector set and color feature vector collection, and truly substance markers image are obtained using default various visual angles sub-space learning model
Its corresponding dimensionality reduction mapping matrix.
Second acquisition module is configurable to, according to each dimensionality reduction mapping matrix acquired in the first acquisition module, obtain each picture
The low-dimensional feature of vegetarian refreshments.
Sort module is configurable to the low-dimensional feature according to each pixel acquired in the second acquisition module, using default
Grader classifies to each pixel to be sorted；
Wherein, it presets shown in various visual angles sub-space learning model such as formula (1).
Further, in this embodiment polarimetric SAR image Supervised classification device can also include mark module；Identify mould
Block is configurable to obtain pole to being identified per a kind of sample in the classification results acquired in sort module with different colours
Change SAR image classification results coloured picture.
Further, in this embodiment A modules can also include filter module, polarization characteristic vector set acquisition module, line
Manage set of eigenvectors acquisition module, color feature vector collection acquisition module and third acquisition module.
Filter module is configurable to be filtered polarimetric SAR image to be sorted using filter.
Polarization characteristic vector set acquisition module is configurable to according to the filtered polarization SAR figure acquired in filter module
Picture obtains polarization characteristic vector set using polarimetric SAR image and goal decomposition method.
Texture feature vector collection acquisition module is configurable to according to the filtered polarization SAR figure acquired in filter module
Picture obtains texture feature vector collection using gray co-occurrence matrix and Gabor filtering methods.
Color feature vector collection acquisition module is configurable to according to the filtered polarization SAR figure acquired in filter module
Picture obtains color feature vector collection using RGB and hsv color histogram method.
Third acquisition module is configurable to according to the polarization characteristic vector acquired in polarization characteristic vector set acquisition module
Collect, the texture feature vector collection and color feature vector collection acquisition module acquired in texture feature vector collection acquisition module are obtained
The color feature vector collection taken obtains the polarization characteristic vector set, texture feature vector collection and color feature vector of training sample
Collection.
The embodiment of above-mentioned polarimetric SAR image supervised classification method, technical principle, it is solved the technical issues of and production
Raw technique effect is similar, and person of ordinary skill in the field can be understood that, for convenience and simplicity of description, on
The specific work process of the polarimetric SAR image Supervised classification device of description and related explanation are stated, aforementioned polarization SAR can be referred to
Image supervised classification method, details are not described herein.
It will be understood by those skilled in the art that above-mentioned polarimetric SAR image supervised classification method further includes some other public affairs
Know structure, such as processor, controller, memory etc., wherein memory includes but not limited to random access memory, flash memory, read-only
Memory, programmable read only memory, volatile memory, nonvolatile memory, serial storage, parallel storage are posted
Storage etc., processor include but not limited to CPLD/FPGA, DSP, arm processor, MIPS processors etc., for unnecessarily mould
Embodiment of the disclosure is pasted, these well known structures are not shown.
It will be understood by those skilled in the art that can adaptively be changed to the module in the device in embodiment
And they are arranged in the one or more devices different from the embodiment.Can in embodiment module or unit or
Component is combined into a module or unit or component, and can be divided into multiple submodule or subelement or subgroup in addition
Part.Other than such feature and/or at least some of process or unit exclude each other, any combinations may be used
To all features disclosed in this specification (including adjoint claim, abstract and attached drawing) and such disclosed any side
All processes or unit of method or equipment are combined.Unless expressly stated otherwise, this specification (including want by adjoint right
Ask, make a summary and attached drawing) disclosed in each feature can be replaced by providing the alternative features of identical, equivalent or similar purpose.
Based on the embodiment of above-mentioned polarimetric SAR image supervised classification method, the present invention also provides a kind of storage devices.
A plurality of program is stored in the present embodiment in storage device, which is suitable for being loaded by processor and being executed to realize above-mentioned pole
Change SAR image supervised classification method.
Based on the embodiment of above-mentioned polarimetric SAR image supervised classification method, the present invention also provides a kind of processing units.
Processing unit may include processor and storage device in the present embodiment.Wherein, processor is adapted for carrying out each program, and storage is set
It is standby to be suitable for storing a plurality of program, and these programs are suitable for being loaded by processor and being executed to realize that above-mentioned polarimetric SAR image has
Supervised classification method.
Person of ordinary skill in the field can be understood that, for convenience and simplicity of description, foregoing description
Storage device, the specific work process of processing unit and related explanation can refer to aforementioned polarimetric SAR image Supervised classification side
Corresponding process in method embodiment, details are not described herein.
The all parts embodiment of the present invention can be with hardware realization, or to run on one or more processors
Software module realize, or realized with combination thereof.It will be understood by those of skill in the art that can use in practice
Microprocessor or digital signal processor (DSP) realize some in server according to the ... of the embodiment of the present invention, client
Or some or all functions of whole components.The present invention is also implemented as one for executing method as described herein
Partly or completely equipment or program of device (for example, PC programs and PC program products).Such journey for realizing the present invention
Sequence can be stored on PC readable mediums, or can be with the form of one or more signal.Such signal can be from
It downloads and obtains on internet website, either provide on carrier signal or provide in any other forms.
In addition, it will be appreciated by those of skill in the art that although some embodiments described herein include other embodiments
In included certain features rather than other feature, but the combination of the feature of different embodiments means in of the invention
Within the scope of and form different embodiments.For example, in claims of the present invention, embodiment claimed
It is one of arbitrary mode to use in any combination.
It should be noted that the present invention will be described rather than limits the invention for above-described embodiment, and ability
Field technique personnel can design alternative embodiment without departing from the scope of the appended claims.In the claims,
Any reference mark between bracket should not be configured to limitations on claims.Word "comprising" does not exclude the presence of not
Element or step listed in the claims.Word "a" or "an" before element does not exclude the presence of multiple such
Element.The present invention can be realized by means of including the hardware of several different elements and by means of properly programmed PC.
If in the unit claim for listing equipment for drying, several in these devices can be by the same hardware branch come specific
It embodies.The use of word first, second, and third does not indicate that any sequence.These words can be construed to title.
So far, it has been combined preferred embodiment shown in the drawings and describes technical scheme of the present invention, still, this field
Technical staff is it is easily understood that protection scope of the present invention is expressly not limited to these specific implementation modes.Without departing from this
Under the premise of the principle of invention, those skilled in the art can make the relevant technologies feature equivalent change or replacement, these
Technical solution after change or replacement is fallen within protection scope of the present invention.
Claims (10)
1. a kind of polarimetric SAR image supervised classification method, which is characterized in that the method includes：
According to polarization characteristic vector set, texture feature vector collection and the color feature vector collection for presetting training sample, and preset
Truly substance markers image obtains its corresponding dimensionality reduction mapping matrix using default various visual angles sub-space learning model；
According to acquired each dimensionality reduction mapping matrix, the low-dimensional feature of each pixel is obtained；
The low-dimensional feature of each pixel according to the acquisition divides each pixel to be sorted using default grader
Class；
Wherein, the default various visual angles sub-space learning model is foundation various visual angles canonical correlation analysis MCCA method models, and
The model built in conjunction with the holding and adaptive weighting parameter of data structure and discriminant information：
Wherein,AndFor the sample data at V visual angle
Collection, it is describedIt is described for the corresponding map vector in V visual angleIt is described for the corresponding visual angle weight at V visual angle
DiFor the dimension of i-th of visual angle characteristic, the n is number of samples, the wiIndicate mapping matrix WiAny one column vector,
AndThe d is the dimension to lower, and the T indicates transposition；
The r and β is parameter.
2. polarimetric SAR image supervised classification method according to claim 1, which is characterized in that the default trained sample
This is the classification according to default truly substance markers, and all kinds of of preset quantity are randomly selected from polarimetric SAR image to be sorted
Other pixel.
3. polarimetric SAR image supervised classification method according to claim 1, which is characterized in that the method is in " foundation
The low-dimensional feature of each pixel classifies to pixel to be sorted using grader " the step of after further include following steps
Suddenly：
Every a kind of sample in polarimetric SAR image in classification results is identified with different colours, obtains Classification of Polarimetric SAR Image result
Coloured picture.
4. polarimetric SAR image supervised classification method according to claim 1, which is characterized in that the method is in " foundation
Polarization characteristic vector set, texture feature vector collection and the color feature vector collection of training sample, utilize various visual angles sub-space learning
Model obtains its corresponding dimensionality reduction mapping matrix " the step of before further include following step：
The polarimetric SAR image to be sorted is filtered using filter；
Pole is obtained using polarization SAR initial data and goal decomposition method according to filtered polarimetric SAR image to be sorted
Change set of eigenvectors；
Texture is obtained using gray co-occurrence matrix and Gabor filtering methods according to filtered polarimetric SAR image to be sorted
Set of eigenvectors；
According to filtered polarimetric SAR image to be sorted, using RGB and hsv color histogram method, obtain color characteristic to
Quantity set；
The training sample is obtained according to acquired polarization characteristic vector set, texture feature vector collection and color feature vector collection
Polarization characteristic vector set, texture feature vector collection and color feature vector collection.
5. polarimetric SAR image supervised classification method according to any one of claims 1-4, which is characterized in that described
Method further includes according to the default various visual angles sub-space learning model, and iterative solution obtains the corresponding mapping matrix in V visual angleWith the corresponding visual angle weight at V visual angleThe specific steps are：
Described in fixationUsing Lagrange multiplier algorithm, the corresponding visual angle weight at V visual angle is obtained
Described in fixationUsing Lagrange multiplier algorithm, the corresponding d map vector w in each visual angle is obtainedi1,
wi2,…,wid, and then constitute mapping matrix
Wi=[wi1,wi2,…,wid]
Wherein, i=1 ..., v.
6. a kind of polarimetric SAR image Supervised classification device, which is characterized in that described device includes the first acquisition module, second
Acquisition module and sort module；
First acquisition module, be configured to according to preset the polarization characteristic vector set of training sample, texture feature vector collection and
It is respectively right to obtain it using default various visual angles sub-space learning model for color feature vector collection, and truly substance markers image
The dimensionality reduction mapping matrix answered；
Second acquisition module is configured to, according to each dimensionality reduction mapping matrix acquired in first acquisition module, obtain each
The low-dimensional feature of pixel；
The sort module is configured to the low-dimensional feature according to each pixel acquired in second acquisition module, using pre-
If grader classifies to each pixel to be sorted；
Wherein, the default various visual angles sub-space learning model is shown below：
7. polarimetric SAR image Supervised classification device according to claim 5, which is characterized in that described device further includes mark
Know module；The mark module is configured to different colours to every a kind of in the classification results acquired in the sort module
Sample is identified, and obtains Classification of Polarimetric SAR Image result coloured picture.
8. polarimetric SAR image Supervised classification device described according to claim 6 or 7, which is characterized in that described device further includes
Filter module, polarization characteristic vector set acquisition module, texture feature vector collection acquisition module, color feature vector collection acquisition module
With third acquisition module；
The filter module is configured to be filtered polarimetric SAR image to be sorted using filter；
The polarization characteristic vector set acquisition module is configured to according to the filtered polarization SAR acquired in the filter module
Image obtains polarization characteristic vector set using polarization SAR initial data and goal decomposition method；
The texture feature vector collection acquisition module is configured to according to the filtered polarization SAR acquired in the filter module
Image obtains texture feature vector collection using gray co-occurrence matrix and Gabor filtering methods；
The color feature vector collection acquisition module is configured to according to the filtered polarization SAR acquired in the filter module
Image obtains color feature vector collection using RGB and hsv color histogram method；
The third acquisition module is configured to according to the polarization characteristic vector acquired in the polarization characteristic vector set acquisition module
Collect, the texture feature vector collection and the color feature vector collection acquired in the texture feature vector collection acquisition module obtain
Color feature vector collection acquired in module, obtain the polarization characteristic vector set of the training sample, texture feature vector collection and
Color feature vector collection.
9. a kind of storage device, wherein being stored with a plurality of program, which is characterized in that described program is suitable for by processor load simultaneously
It executes to realize claim 1-5 any one of them polarimetric SAR image supervised classification methods.
10. a kind of processing unit, including
Processor is adapted for carrying out each program；And
Storage device is suitable for storing a plurality of program；
It is characterized in that, described program is suitable for being loaded by processor and being executed to realize：Claim 1-5 any one of them pole
Change SAR image supervised classification method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810036297.5A CN108460402B (en) | 2018-01-15 | 2018-01-15 | Method and device for supervised classification of polarized SAR (synthetic aperture radar) images |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810036297.5A CN108460402B (en) | 2018-01-15 | 2018-01-15 | Method and device for supervised classification of polarized SAR (synthetic aperture radar) images |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108460402A true CN108460402A (en) | 2018-08-28 |
| CN108460402B CN108460402B (en) | 2020-07-24 |
Family
ID=63221456
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810036297.5A Active CN108460402B (en) | 2018-01-15 | 2018-01-15 | Method and device for supervised classification of polarized SAR (synthetic aperture radar) images |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108460402B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110110579A (en) * | 2019-02-21 | 2019-08-09 | 中国科学院电子学研究所 | A kind of Classification of Polarimetric SAR Image method, apparatus and storage medium |
| CN111046933A (en) * | 2019-12-03 | 2020-04-21 | 东软集团股份有限公司 | Image classification method and device, storage medium and electronic equipment |
| CN111079666A (en) * | 2019-12-20 | 2020-04-28 | 广州市鑫广飞信息科技有限公司 | Ground object identification method, device, equipment and storage medium |
| CN113191179A (en) * | 2020-12-21 | 2021-07-30 | 广州蓝图地理信息技术有限公司 | Remote sensing image classification method based on gray level co-occurrence matrix and BP neural network |
| CN114411337A (en) * | 2022-01-21 | 2022-04-29 | 山东晶创新材料科技有限公司 | Preparation method of ultra-wide polypropylene filament non-woven fabric |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101635027A (en) * | 2009-09-03 | 2010-01-27 | 西安交通大学 | En-ULLELDA-based method of multi-view model recognition |
| CN105718915A (en) * | 2016-01-27 | 2016-06-29 | 中国石油大学(华东) | Face identification method and system based on multi-visual-angle typical correlation analysis |
| CN105825223A (en) * | 2016-03-09 | 2016-08-03 | 西安电子科技大学 | Polarization SAR terrain classification method based on deep learning and distance metric learning |
| CN105956611A (en) * | 2016-04-25 | 2016-09-21 | 西安电子科技大学 | SAR image target identification method based on authentication non-linear dictionary learning |
-
2018
- 2018-01-15 CN CN201810036297.5A patent/CN108460402B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101635027A (en) * | 2009-09-03 | 2010-01-27 | 西安交通大学 | En-ULLELDA-based method of multi-view model recognition |
| CN105718915A (en) * | 2016-01-27 | 2016-06-29 | 中国石油大学(华东) | Face identification method and system based on multi-visual-angle typical correlation analysis |
| CN105825223A (en) * | 2016-03-09 | 2016-08-03 | 西安电子科技大学 | Polarization SAR terrain classification method based on deep learning and distance metric learning |
| CN105956611A (en) * | 2016-04-25 | 2016-09-21 | 西安电子科技大学 | SAR image target identification method based on authentication non-linear dictionary learning |
Non-Patent Citations (2)
| Title |
|---|
| MEINA KAN等: "Multi-view Discriminant Analysis", 《IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE》 * |
| XIAYUAN HUANG等: "Local Discriminant Canonical Correlation Analysis for Supervised POLSAR", 《IEEE GEOSCIENCE AND REMOTE SENSING LETTERS》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110110579A (en) * | 2019-02-21 | 2019-08-09 | 中国科学院电子学研究所 | A kind of Classification of Polarimetric SAR Image method, apparatus and storage medium |
| CN111046933A (en) * | 2019-12-03 | 2020-04-21 | 东软集团股份有限公司 | Image classification method and device, storage medium and electronic equipment |
| CN111046933B (en) * | 2019-12-03 | 2024-03-05 | 东软集团股份有限公司 | Image classification method, device, storage medium and electronic equipment |
| CN111079666A (en) * | 2019-12-20 | 2020-04-28 | 广州市鑫广飞信息科技有限公司 | Ground object identification method, device, equipment and storage medium |
| CN111079666B (en) * | 2019-12-20 | 2024-03-19 | 广州市鑫广飞信息科技有限公司 | Ground object identification method, device, equipment and storage medium |
| CN113191179A (en) * | 2020-12-21 | 2021-07-30 | 广州蓝图地理信息技术有限公司 | Remote sensing image classification method based on gray level co-occurrence matrix and BP neural network |
| CN114411337A (en) * | 2022-01-21 | 2022-04-29 | 山东晶创新材料科技有限公司 | Preparation method of ultra-wide polypropylene filament non-woven fabric |
| CN114411337B (en) * | 2022-01-21 | 2023-03-14 | 山东晶创新材料科技有限公司 | Preparation method of ultra-wide polypropylene filament non-woven fabric |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108460402B (en) | 2020-07-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108460402A (en) | Polarimetric SAR image supervised classification method and device | |
| CN105260729B (en) | A kind of satellite remote-sensing image cloud amount computational methods based on random forest | |
| Laliberte et al. | A comparison of three feature selection methods for object-based classification of sub-decimeter resolution UltraCam-L imagery | |
| Al Bashish et al. | A framework for detection and classification of plant leaf and stem diseases | |
| Loosvelt et al. | Random Forests as a tool for estimating uncertainty at pixel-level in SAR image classification | |
| Jaynes et al. | Cluster analysis of spatiotemporal corn yield patterns in an Iowa field | |
| Du et al. | Spatial and spectral unmixing using the beta compositional model | |
| Mishra et al. | Land cover classification of PALSAR images by knowledge based decision tree classifier and supervised classifiers based on SAR observables | |
| DE60221109T2 (en) | IMAGE SEARCH USING A DISTANCE | |
| CN103413151B (en) | Hyperspectral image classification method based on figure canonical low-rank representation Dimensionality Reduction | |
| CN105718942B (en) | High spectrum image imbalance classification method based on average drifting and over-sampling | |
| CN105760900B (en) | Hyperspectral image classification method based on neighbour's propagation clustering and sparse Multiple Kernel Learning | |
| CN107229917A (en) | A kind of several remote sensing image general character well-marked target detection methods clustered based on iteration | |
| CN107679509A (en) | A kind of small ring algae recognition methods and device | |
| CN106897741B (en) | In conjunction with the polarization SAR terrain classification method of rotational domain polarization coherence property | |
| CN105913092B (en) | Figure canonical hyperspectral image band selection method based on sub-space learning | |
| CN104700116B (en) | The sorting technique of the Polarimetric SAR Image atural object represented based on multi-layer quantum ridge ripple | |
| CN109840542B (en) | Self-adaptive dimension decision tree classification method based on polarization characteristics | |
| Zeyada et al. | Evaluation of the discrimination capability of full polarimetric SAR data for crop classification | |
| Florindo et al. | Brachiaria species identification using imaging techniques based on fractal descriptors | |
| CN107832786A (en) | A kind of recognition of face sorting technique based on dictionary learning | |
| CN105096293B (en) | Method and apparatus for handling the block to be processed of sediment urinalysis image | |
| CN105138966A (en) | Quick density peak value clustering based polarimetric SAR image classification method | |
| Yalcin | Phenology monitoring of agricultural plants using texture analysis | |
| Lambert et al. | Testing the ability of unmanned aerial systems and machine learning to map weeds at subfield scales: a test with the weed Alopecurus myosuroides (Huds) |
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 |