CN104376335A - Semi-supervised hyperspectral remote sensing image classification method based on information entropies - Google Patents
Semi-supervised hyperspectral remote sensing image classification method based on information entropies Download PDFInfo
- Publication number
- CN104376335A CN104376335A CN201410668553.4A CN201410668553A CN104376335A CN 104376335 A CN104376335 A CN 104376335A CN 201410668553 A CN201410668553 A CN 201410668553A CN 104376335 A CN104376335 A CN 104376335A
- Authority
- CN
- China
- Prior art keywords
- remote sensing
- classification
- pixel
- hyperspectral
- semi
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
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
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
- G06F18/2155—Generating training patterns; Bootstrap methods, e.g. bagging or boosting characterised by the incorporation of unlabelled data, e.g. multiple instance learning [MIL], semi-supervised techniques using expectation-maximisation [EM] or naïve labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/19—Recognition using electronic means
- G06V30/192—Recognition using electronic means using simultaneous comparisons or correlations of the image signals with a plurality of references
- G06V30/194—References adjustable by an adaptive method, e.g. learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Databases & Information Systems (AREA)
- Multimedia (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a semi-supervised hyperspectral remote sensing image classification method based on information entropies and relates to the field of remote sensing. The method specifically comprises the steps of (1) inputting a hyperspectral remote sensing image, (2) inputting a training sample set, (3) inputting a category set corresponding to training samples, (4) calculating the probability of the category which each image element in the hyperspectral remote sensing image represents through a multi-classification linear regression method, (5) outputting the categories corresponding to all the image elements according to the calculated probabilities of all the image elements, (6) outputting the classification result and judging the accuracy of the output result, (7) converting the probabilities of all the image elements in the remote sensing image into indeterminacy through the Renyi entropy algorithm, (8) converting unmarked label image elements in the hyperspectral remote sensing image into marked label image elements according to the indeterminacy, (9) adding new marked labels into the training set, and (10) conducting iteration operation. The hyperspectral remote sensing image classification method has the advantages of being easy to realize, low in calculation complexity and the like.
Description
Technical field
What the present invention relates to is technical field of remote sensing image processing, be specially adapted to target in hyperspectral remotely sensed image, propose a kind of semi-supervised Hyperspectral Remote Sensing Imagery Classification method based on information entropy, can be applicable in Hyperspectral Remote Sensing Imagery Classification and the multiple application such as identification and Land Use/Cover Classification.
Background technology
The research of Hyperspectral Remote Sensing Imagery Classification method is that airborne-remote sensing disposal route is most crucial, one of the research field of most critical, its general objective is divided into dissimilar by pixels all in remote sensing image, clearly can reflect the space distribution of certain class and the details of all kinds of atural object thus, be all obtained at military or civil area to apply widely.But because high, the data volume of its dimension is large, data ambiguity, be subject to the factors such as Hughes phenomena impair and jointly cause the difficulty that terrestrial object information automatic classification extracts, become one of bottleneck of restriction high-spectrum remote-sensing application.
Pattern-recognition and artificial intelligence obtain development at full speed in last century, and are used widely at remote sensing fields, have occurred many effective intelligent method for classifying, provide powerful support for solving terrestrial object information automatic classification.Hyperspectral classification method common is at present divided into supervised classification and Non-surveillance clustering method two kinds, but these two kinds of methods still exist some insoluble problems in high spectrum image actual classification.The subject matter of measure of supervision is the quality and scale that learning process depends on training set to a great extent, because the cost of mark training sample classification is very high, training dataset obtains cost and causes more greatly training sample quantity very limited, and the acquisition of sample more difficult work often in hyperspectral classification, particularly adopt during high dimensional feature vector and require that the sample number of every class is all high than intrinsic dimensionality, the precision therefore in high dimensional information process and efficiency and just create great contradiction between the meticulous spectrum of high-spectrum remote-sensing information and big data quantity.Non-supervisory method subject matter only uses data untagged, and the classification that can produce cluster is uncertain, and nicety of grading is low, cannot reach the classifying quality of expection.
Semi-supervised learning (semi-supervised learning) sorting technique of development in recent years is a kind of new machine learning method utilizing the sample do not marked to carry out training classifier, there is small-sample learning, noiseproof feature be strong, learning efficiency is high, the training features such as cost is little, become the new study hotspot of pattern-recognition and one, machine learning field.Therefore, along with the develop rapidly of imaging spectral technology and the magnanimity of Hyperspectral imaging are emerged in large numbers, the research theoretical method of semi-supervised learning and the application in Hyperspectral Image Classification, to the classification performance improving target in hyperspectral remotely sensed image, improve precision and the efficiency of Hyperspectral imaging process, realize making full use of of Hyperspectral imaging and there is great theory value and practical application meaning.
There are some target in hyperspectral remotely sensed image semisupervised classification methods present stage, a kind of based on the semi-supervised sparse Hyperspectral Remote Sensing Imagery Classification method (Intellectual Property Right Bureau of the RPC application publication number CN103593676A) differentiating to embed its to adopt semi-supervised sparse discriminating embedded mobile GIS to carry out dimension to target in hyperspectral remotely sensed image brief, combine neighbour's manifold structure and openness advantage, sparse Remodeling not only between retain sample, utilize simultaneously and have mark training sample and part Using Non-labeled Training Sample to find to contain in the inherent attribute of high dimensional data and low dimensional manifold structure on a small quantity, the nicety of grading to atural object classification in target in hyperspectral remotely sensed image can be improved.And based on using class probability metrics to calculate the Neighbor Points of sample point in the semi-supervised high-spectral remote sensing sorting technique (Intellectual Property Right Bureau of the RPC application publication number CN102096825A) of figure, use the semi-supervised Accurate classification realized high spectrum image of deriving, substantially reduce computation complexity.
A kind of semi-supervised Hyperspectral Remote Sensing Imagery Classification method based on information entropy that the present invention proposes, for the feature of high-spectral data, on the basis choosing less markd training label, by logistic algorithm of classifying more, dope preliminary classification result, then weighed the energy of image by renyi entropy, choose the pixel comprising information maximization and add in training sample, carry out prediction classification again, finally realize the object of target in hyperspectral remotely sensed image Accurate classification.New method has good classifying quality, and overall classification accuracy is high, and mistake is divided and leakage divides mistake less, and the geographic classification intensive for multiclass has advantage more.
Summary of the invention
For the deficiency that prior art exists, the present invention seeks to be to provide a kind of semi-supervised Hyperspectral Remote Sensing Imagery Classification method based on information entropy, solve a large amount of unmarked label existed in target in hyperspectral remotely sensed image and utilize insufficient problem, and improve precision and the effect of classification.
To achieve these goals, the present invention realizes by the following technical solutions: a kind of semi-supervised Hyperspectral Remote Sensing Imagery Classification method based on information entropy, and it comprises the following steps:
(1) target in hyperspectral remotely sensed image data are inputted;
(2) the training sample set of the middle target in hyperspectral remotely sensed image data of input step (1);
(3) the class categories collection in input step (2) corresponding to training sample set;
(4) rely on training sample set and class categories collection, use the method for many classification linear regressions to calculate the probability of the classification in target in hyperspectral remotely sensed image representated by each pixel;
(5) each pixel class probability calculated according to step (4) exports classification corresponding to each pixel;
(6) this classification results is exported, and judge the precision of Output rusults, if iteration then carries out step (7) for the first time, otherwise compare with a front Output rusults, the threshold value differing by more than setting in advance when both then proceeds step (7), and the threshold value being less than setting else if then exports net result;
(7) probability of Renyi entropy algorithm to pixel each in remote sensing image is utilized to be converted to uncertainty;
(8) according to uncertainty, label pixel unmarked in Hyperspectral imaging is converted to markup tags pixel;
(9) new markup tags is joined in training set;
(10) return step (4) iteration to run.
Beneficial effect of the present invention: on the basis choosing less markd training label, by logistic algorithm of classifying more, dope preliminary classification result, then the energy of image is weighed by renyi entropy, choosing the pixel comprising information maximization adds in training sample, carry out prediction classification again, finally realize the object of target in hyperspectral remotely sensed image Accurate classification.
Accompanying drawing explanation
The present invention is described in detail below in conjunction with the drawings and specific embodiments;
Fig. 1 is method flow diagram of the present invention;
Fig. 2 is that the classification that high-spectrum remote-sensing original image, object spectrum figure and each atural object that the present invention uses are corresponding indicates;
Fig. 3 is the final classifying quality figure of each algorithm of using in the present invention.
Embodiment
The technological means realized for making the present invention, creation characteristic, reaching object and effect is easy to understand, below in conjunction with embodiment, setting forth the present invention further.
With reference to Fig. 1, this embodiment is by the following technical solutions: a kind of semi-supervised Hyperspectral Remote Sensing Imagery Classification method based on information entropy, its concrete steps are: (1) input target in hyperspectral remotely sensed image data: input target in hyperspectral remotely sensed image type is not specific, the target in hyperspectral remotely sensed image that spaceborne imaging spectrometer obtains and the target in hyperspectral remotely sensed image that airborne imaging spectrum instrument obtains.
(2) training sample set is inputted: the training sample set of input is the target in hyperspectral remotely sensed image inputted for step (1), through the training sample of expertise and Knowledge Acquirement.
(3) input classification collection corresponding to training sample: the classification collection that input training sample is corresponding is the target in hyperspectral remotely sensed image inputted for step (1), through the classification collection of expertise and Knowledge Acquirement, namely final image is divided into a few class.
(4) use the method for many classification linear regressions to calculate the probability of the classification in target in hyperspectral remotely sensed image representated by each pixel: to rely on training sample set and class categories collection, it is as follows that the probabilistic method of the classification using the method for many classification linear regressions to calculate in target in hyperspectral remotely sensed image representated by each pixel employs the concrete formula of method that many classification logistic return:
Formula (11) represents the new probability formula x of the probability of generation event many classification Logistic model result variable
irepresent the DN value of a certain pixel, C represents classification collection.
H (x
i) solution formula is such as formula shown in (12)
h(x
i)＝β
0+β
1x
i1+β
2x
i2…β
px
ip+ε (12)
Estimation coefficient β is that formula 13 gives solution by the maximum estimation criterion posteriority of Bayes and log-likelihood algorithm estimation coefficient
X in above formula
irepresent certain pixel point in training set, t (x
i) represent the DN value of time pixel point, can calculated related coefficient thus the probable value of each point can be calculated according to training set.
(5) each pixel class probability calculated exports classification corresponding to each pixel: the classification that each pixel class probability calculated exports each pixel corresponding refers to:
(6) this classification results is exported, and judge the precision of Output rusults, if iteration then carries out step below for the first time, otherwise compare with a front Output rusults, step below the threshold value differing by more than in advance setting when both then proceeds, the threshold value being less than setting else if then exports net result;
(7) probability of Renyi entropy algorithm to pixel each in remote sensing image is utilized to be converted to uncertainty: to utilize the probability of Renyi entropy algorithm to pixel each in remote sensing image to be converted to uncertain formula and refer to:
The homogeneity of Renyi entropy reflection property value distribution, and it is the information of a scale measurement and probabilistic degree of quantity of information.Wherein p (x
i) be the probable value of target in hyperspectral remotely sensed image pixel point.Generally the value of α is taken as 2, and institute with the formula (15) can be converted to formula (16).
When calculating the Renyi entropy of airborne-remote sensing pixel, there will be the situation of maximum value, in order to make net result according to accurate, this method have employed normalized, as shown in formula (17).
(8) according to uncertainty, label pixel unmarked in Hyperspectral imaging is converted to markup tags pixel: in Hyperspectral imaging, unmarked label pixel is converted to markup tags pixel and refers to according to information entropy theory, uncertain maximum element information amount is maximum.
(9) new markup tags is joined in training set: new markup tags is joined in training set and refers to that selecting the maximum pixel of Renyi entropy to add in training set forms new training set.
(10) iteration is run.
Embodiment 1: based on the semi-supervised Hyperspectral Remote Sensing Imagery Classification method of information entropy with the step of above-mentioned embodiment; Fig. 2 (a) is the high-spectrum remote-sensing original image that the present invention uses, it is the image in Florida State Kennedy Sapce Centre (KSC) on the March 23rd, 1996 obtained by the airborne imaging spectrum instrument AVIRIS of US National Aeronautics and Space Administration (NASA), one has 224 wave bands, spectral range 400-2500, spectral resolution 10nm, spatial resolution 18m, for the data of study area, eliminate the wave band of water vapor absorption impact and low SNR, select 120 wave bands to analyze altogether.Training data is that the image provided according to Landsat thematic mapper (Landsat Thematic Mapper) carries out selecting, and according to the interpretation to image, land cover pattern in this region is divided into the classification that 13 large.Fig. 2 (b) represents the spectral curve of 13 classification atural objects.Fig. 2 (c) represents the legend of 13 classifications.
In order to verify the validity of the inventive method, and under identical sample conditions, the semi-supervised Hyperspectral Remote Sensing Imagery Classification method that the present invention is based on information entropy is compared with representative machine learning classification algorithm K-means algorithm, minimum distance method classification (Minimum Distance) algorithm, support vector cassification (support vector machine algorithm) algorithm.Classification results demonstrates, k-means sorting algorithm, although Riming time of algorithm is the shortest, error rate is higher, and class categories is specified by sorter, leaks and divides rate higher.Minimum distance method classification (MinimumDistance) Riming time of algorithm is moderate, but leaks classification error relatively seriously, and overall classification accuracy (Overall Accuracy) is low, even if trace is also difficult to find in the markd place of training sample.The quality comparation of support vector cassification (support vector machine algorithm) algorithm is high, travelling speed is fast, overall classification accuracy (Overall Accuracy) is higher, best to large stretch of continuum classifying quality, but two of support vector cassification algorithm parameter C and ó, to constantly change according to different pieces of information parameter, need constantly to regulate reply data sensitive, particularly low to the analysis ability of tiny atural object, and the inventive method does not need regulating parameter, easy to use.Fig. 3 (a) is the result figure of k-means sorting algorithm, Fig. 3 (b) is the result figure of minimum distance method sorting algorithm, the result figure of Fig. 3 (c) to be the result figure of support vector cassification algorithm, Fig. 3 (d) be sorting algorithm that the present invention proposes.Table 1 lists the final nicety of grading of various different sorting technique in this experiment.
Table one
Sorting technique of the present invention can find the uncertain information be hidden in high-spectral data, according to the theory of information entropy, a small amount of having can be made full use of and mark training sample and a large amount of unmarked training sample, compare additive method and divide the classification results figure obtained better effects if, nicety of grading is significantly improved.
Along with the development of high spectrum resolution remote sensing technique, the approach obtaining target in hyperspectral remotely sensed image increases greatly, and can be more and more easier, and thing followed application also can get more and more, and will relate to numerous fields.Hyperspectral Remote Sensing Imagery Classification technology is one of Focal point and difficult point in target in hyperspectral remotely sensed image preconditioning technique, therefore study Hyperspectral Remote Sensing Imagery Classification method and have important realistic meaning, the present invention is that the development of Hyperspectral Remote Sensing Imagery Classification method provides a kind of new thinking.
More than show and describe ultimate principle of the present invention and principal character and advantage of the present invention.The technician of the industry should understand; the present invention is not restricted to the described embodiments; what describe in above-described embodiment and instructions just illustrates principle of the present invention; without departing from the spirit and scope of the present invention; the present invention also has various changes and modifications, and these changes and improvements all fall in the claimed scope of the invention.Application claims protection domain is defined by appending claims and equivalent thereof.
Claims (9)
1. based on a semi-supervised Hyperspectral Remote Sensing Imagery Classification method for information entropy, it is characterized in that, comprise the following steps: (1) input target in hyperspectral remotely sensed image data;
(2) training sample set of the middle target in hyperspectral remotely sensed image data of input step (1);
(3) the class categories collection in input step (2) corresponding to training sample set;
(4) rely on training sample set and class categories collection, use the method for many classification linear regressions to calculate the probability of the classification in target in hyperspectral remotely sensed image representated by each pixel;
(5) each pixel class probability calculated according to step (4) exports classification corresponding to each pixel;
(6) this classification results is exported, and judge the precision of Output rusults, if iteration then carries out step (7) for the first time, otherwise compare with a front Output rusults, the threshold value differing by more than setting in advance when both then proceeds step (7), and the threshold value being less than setting else if then exports net result;
(7) probability of Renyi entropy algorithm to pixel each in remote sensing image is utilized to be converted to uncertainty;
(8) according to uncertainty, label pixel unmarked in Hyperspectral imaging is converted to markup tags pixel;
(9) new markup tags is joined in training set;
(10) return step (4) iteration to run.
2. the semi-supervised Hyperspectral Remote Sensing Imagery Classification method based on information entropy according to claim 1, it is characterized in that: in described step (1), input target in hyperspectral remotely sensed image type is not specific, the target in hyperspectral remotely sensed image that spaceborne imaging spectrometer obtains and the target in hyperspectral remotely sensed image that airborne imaging spectrum instrument obtains.
3. the semi-supervised Hyperspectral Remote Sensing Imagery Classification method based on information entropy according to claim 1, it is characterized in that: the training sample set of input is the target in hyperspectral remotely sensed image inputted for step (1) in described step (2), through the training sample of expertise and Knowledge Acquirement.
4. the semi-supervised Hyperspectral Remote Sensing Imagery Classification method based on information entropy according to claim 1, it is characterized in that: the classification collection that in described step (3), input training sample is corresponding is the target in hyperspectral remotely sensed image inputted for step (1), through the classification collection of expertise and Knowledge Acquirement, namely final image is divided into a few class.
5. the semi-supervised Hyperspectral Remote Sensing Imagery Classification method based on information entropy according to claim 1, it is characterized in that: described step (4) dependence training sample set and class categories collection, the concrete formula of method that the probabilistic method of the classification using the method for classifying linear regression to calculate in target in hyperspectral remotely sensed image representated by each pixel employs logistic recurrence of classifying more more is as follows:
Formula (11) represents the new probability formula x of the probability of generation event many classification Logistic model result variable
irepresent the DN value of a certain pixel, C represents classification collection;
H (x
i) solution formula is such as formula shown in (12)
h(x
i)＝β
0+β
1x
i1+β
2x
i2...β
px
ip+ε (12)
Estimation coefficient β is that formula 13 gives solution by the maximum estimation criterion posteriority of Bayes and log-likelihood algorithm estimation coefficient
X in above formula
irepresent certain pixel point in training set, t (x
i) represent the DN value of time pixel point, can calculated related coefficient thus the probable value of each point can be calculated according to training set.
6. the semi-supervised Hyperspectral Remote Sensing Imagery Classification method based on information entropy according to claim 1, is characterized in that: the classification that each pixel class probability calculated in described step (5) exports each pixel corresponding refers to:
。
7. the semi-supervised Hyperspectral Remote Sensing Imagery Classification method based on information entropy according to claim 1, it is characterized in that: described step exports this classification results in (6), and judge the precision of Output rusults, if iteration then carries out step below for the first time, otherwise compare with a front Output rusults, step below the threshold value differing by more than in advance setting when both then proceeds, the threshold value being less than setting else if then exports net result.
8. the semi-supervised Hyperspectral Remote Sensing Imagery Classification method based on information entropy according to claim 1, is characterized in that: utilize the probability of Renyi entropy algorithm to pixel each in remote sensing image to be converted to uncertain formula in described step (7) and refer to:
The homogeneity of Renyi entropy reflection property value distribution, and it is the information of a scale measurement and probabilistic degree of quantity of information.Wherein p (x
i) be the probable value of target in hyperspectral remotely sensed image pixel point.Generally the value of α is taken as 2, and institute with the formula (15) can be converted to formula (16):
When calculating the Renyi entropy of airborne-remote sensing pixel, there will be the situation of maximum value, in order to make net result according to accurate, this method have employed normalized, as shown in formula (17):
。
9. the semi-supervised Hyperspectral Remote Sensing Imagery Classification method based on information entropy according to claim 1, it is characterized in that: in described step (8) and (9), label pixel unmarked in Hyperspectral imaging is converted to markup tags pixel and adds in training set and refer to according to information entropy theory, uncertain maximum element information amount is maximum, and this method is selected the maximum pixel of Renyi entropy to add in training set to form new training set.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410668553.4A CN104376335A (en) | 2014-11-13 | 2014-11-13 | Semi-supervised hyperspectral remote sensing image classification method based on information entropies |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410668553.4A CN104376335A (en) | 2014-11-13 | 2014-11-13 | Semi-supervised hyperspectral remote sensing image classification method based on information entropies |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104376335A true CN104376335A (en) | 2015-02-25 |
Family
ID=52555231
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410668553.4A Pending CN104376335A (en) | 2014-11-13 | 2014-11-13 | Semi-supervised hyperspectral remote sensing image classification method based on information entropies |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104376335A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105320969A (en) * | 2015-11-20 | 2016-02-10 | 北京理工大学 | A heart rate variability feature classification method based on multi-scale Renyi entropy |
| CN107729918A (en) * | 2017-09-15 | 2018-02-23 | 电子科技大学 | Cellular automata based on Cost Sensitive Support Vector Machines emerges in large numbers the sorting technique of phenomenon |
| CN108038448A (en) * | 2017-12-13 | 2018-05-15 | 河南理工大学 | Semi-supervised random forest Hyperspectral Remote Sensing Imagery Classification method based on weighted entropy |
| CN108388828A (en) * | 2017-07-13 | 2018-08-10 | 中国科学院遥感与数字地球研究所 | A kind of seashore wetland land cover pattern information extracting method of comprehensive multi- source Remote Sensing Data data |
| CN109784392A (en) * | 2019-01-07 | 2019-05-21 | 华南理工大学 | A kind of high spectrum image semisupervised classification method based on comprehensive confidence |
| CN110717531A (en) * | 2019-09-27 | 2020-01-21 | 华东师范大学 | Method for detecting classified change type based on uncertainty analysis and Bayesian fusion |
| CN111325771A (en) * | 2020-02-17 | 2020-06-23 | 武汉大学 | High-resolution remote sensing image change detection method based on image fusion framework |
| CN111460943A (en) * | 2020-03-24 | 2020-07-28 | 山西大学 | Remote sensing image ground object classification method and system |
| CN112784818A (en) * | 2021-03-03 | 2021-05-11 | 电子科技大学 | Identification method based on grouping type active learning on optical remote sensing image |
| CN112905823A (en) * | 2021-02-22 | 2021-06-04 | 深圳市国科光谱技术有限公司 | Hyperspectral substance detection and identification system and method based on big data platform |
| WO2023000160A1 (en) * | 2021-07-20 | 2023-01-26 | 海南长光卫星信息技术有限公司 | Hyperspectral remote sensing image semi-supervised classification method, apparatus, and device, and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030012398A1 (en) * | 2001-07-11 | 2003-01-16 | Sunshine Jessica Miriam | Method for selecting representative endmember components from spectral data |
| US20050047663A1 (en) * | 2002-06-28 | 2005-03-03 | Keenan Daniel M. | Spectral mixture process conditioned by spatially-smooth partitioning |
| CN102096825A (en) * | 2011-03-23 | 2011-06-15 | 西安电子科技大学 | Graph-based semi-supervised high-spectral remote sensing image classification method |
| CN103345643A (en) * | 2013-06-13 | 2013-10-09 | 南京信息工程大学 | Remote sensing image classification method |
| CN103593676A (en) * | 2013-11-29 | 2014-02-19 | 重庆大学 | High-spectral remote-sensing image classification method based on semi-supervision sparse discriminant embedding |
-
2014
- 2014-11-13 CN CN201410668553.4A patent/CN104376335A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030012398A1 (en) * | 2001-07-11 | 2003-01-16 | Sunshine Jessica Miriam | Method for selecting representative endmember components from spectral data |
| US20050047663A1 (en) * | 2002-06-28 | 2005-03-03 | Keenan Daniel M. | Spectral mixture process conditioned by spatially-smooth partitioning |
| CN102096825A (en) * | 2011-03-23 | 2011-06-15 | 西安电子科技大学 | Graph-based semi-supervised high-spectral remote sensing image classification method |
| CN103345643A (en) * | 2013-06-13 | 2013-10-09 | 南京信息工程大学 | Remote sensing image classification method |
| CN103593676A (en) * | 2013-11-29 | 2014-02-19 | 重庆大学 | High-spectral remote-sensing image classification method based on semi-supervision sparse discriminant embedding |
Non-Patent Citations (1)
| Title |
|---|
| 王春阳等: ""A New Semi-supervised Classification Method of Hyperspectral Image based on Combining Renyi Entropy and Multinomial Logistic Regression Algorithm"", 《INTERNATIONAL JOURNAL OF SIGNAL PROCESSING,IMAGE PROCESSING AND PATTERN RECOGNITION》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105320969A (en) * | 2015-11-20 | 2016-02-10 | 北京理工大学 | A heart rate variability feature classification method based on multi-scale Renyi entropy |
| CN108388828A (en) * | 2017-07-13 | 2018-08-10 | 中国科学院遥感与数字地球研究所 | A kind of seashore wetland land cover pattern information extracting method of comprehensive multi- source Remote Sensing Data data |
| CN107729918B (en) * | 2017-09-15 | 2024-03-19 | 电子科技大学 | Classification method for cellular automaton emerging phenomenon based on cost-sensitive support vector machine |
| CN107729918A (en) * | 2017-09-15 | 2018-02-23 | 电子科技大学 | Cellular automata based on Cost Sensitive Support Vector Machines emerges in large numbers the sorting technique of phenomenon |
| CN108038448A (en) * | 2017-12-13 | 2018-05-15 | 河南理工大学 | Semi-supervised random forest Hyperspectral Remote Sensing Imagery Classification method based on weighted entropy |
| CN109784392A (en) * | 2019-01-07 | 2019-05-21 | 华南理工大学 | A kind of high spectrum image semisupervised classification method based on comprehensive confidence |
| CN110717531A (en) * | 2019-09-27 | 2020-01-21 | 华东师范大学 | Method for detecting classified change type based on uncertainty analysis and Bayesian fusion |
| CN111325771A (en) * | 2020-02-17 | 2020-06-23 | 武汉大学 | High-resolution remote sensing image change detection method based on image fusion framework |
| CN111325771B (en) * | 2020-02-17 | 2022-02-01 | 武汉大学 | High-resolution remote sensing image change detection method based on image fusion framework |
| CN111460943A (en) * | 2020-03-24 | 2020-07-28 | 山西大学 | Remote sensing image ground object classification method and system |
| CN112905823A (en) * | 2021-02-22 | 2021-06-04 | 深圳市国科光谱技术有限公司 | Hyperspectral substance detection and identification system and method based on big data platform |
| CN112905823B (en) * | 2021-02-22 | 2023-10-31 | 深圳市国科光谱技术有限公司 | Hyperspectral substance detection and identification system and method based on big data platform |
| CN112784818A (en) * | 2021-03-03 | 2021-05-11 | 电子科技大学 | Identification method based on grouping type active learning on optical remote sensing image |
| CN112784818B (en) * | 2021-03-03 | 2023-03-14 | 电子科技大学 | Identification method based on grouping type active learning on optical remote sensing image |
| WO2023000160A1 (en) * | 2021-07-20 | 2023-01-26 | 海南长光卫星信息技术有限公司 | Hyperspectral remote sensing image semi-supervised classification method, apparatus, and device, and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104376335A (en) | Semi-supervised hyperspectral remote sensing image classification method based on information entropies | |
| Sahin et al. | A comparative assessment of canonical correlation forest, random forest, rotation forest and logistic regression methods for landslide susceptibility mapping | |
| CN108038445B (en) | SAR automatic target identification method based on multi-view deep learning framework | |
| Schneider | Monitoring land cover change in urban and peri-urban areas using dense time stacks of Landsat satellite data and a data mining approach | |
| CN102096825B (en) | Graph-based semi-supervised high-spectral remote sensing image classification method | |
| CN102073879B (en) | Method for identifying characteristic land categories of ocean remote sensing images of coast on basis of semi-supervised learning | |
| CN101551856B (en) | SAR target recognition method based on sparse least squares support vector machine | |
| Bishop et al. | Gain form of the ensemble transform Kalman filter and its relevance to satellite data assimilation with model space ensemble covariance localization | |
| Xiao et al. | Classification of weather phenomenon from images by using deep convolutional neural network | |
| Li | Research on radar signal recognition based on automatic machine learning | |
| Zhang et al. | A GANs-based deep learning framework for automatic subsurface object recognition from ground penetrating radar data | |
| CN108229551B (en) | Hyperspectral remote sensing image classification method based on compact dictionary sparse representation | |
| Qing et al. | Spatial‐spectral classification of hyperspectral images: a deep learning framework with Markov Random fields based modelling | |
| CN103942749A (en) | Hyperspectral ground feature classification method based on modified cluster hypothesis and semi-supervised extreme learning machine | |
| Gao et al. | A multidepth and multibranch network for hyperspectral target detection based on band selection | |
| CN105117736A (en) | Polarized SAR image classification method based on sparse depth stack network | |
| CN114821198A (en) | Cross-domain hyperspectral image classification method based on self-supervision and small sample learning | |
| CN103745233A (en) | Hyper-spectral image classifying method based on spatial information transfer | |
| CN103425995A (en) | Hyperspectral image classification method based on area similarity low rank expression dimension reduction | |
| CN103226825B (en) | Based on the method for detecting change of remote sensing image of low-rank sparse model | |
| Pangaliman et al. | Machine learning predictive models for improved acoustic disdrometer | |
| Han et al. | Targets mask U-Net for wind turbines detection in remote sensing images | |
| Silverman et al. | Predicting origins of coherent air mass trajectories using a neural network—the case of dry intrusions | |
| Kiran et al. | Fuzzy ensemble clustering approach to address regionalization uncertainties in flood frequency analysis | |
| Coburn et al. | Projecting future energy production from operating wind farms in North America. Part II: Statistical downscaling |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20150225 |
|
| WD01 | Invention patent application deemed withdrawn after publication |