CN101673339A - Target identification method of high spectroscopic data rearranged based on spectral absorption characteristics - Google Patents
Target identification method of high spectroscopic data rearranged based on spectral absorption characteristics Download PDFInfo
- Publication number
- CN101673339A CN101673339A CN200910093546A CN200910093546A CN101673339A CN 101673339 A CN101673339 A CN 101673339A CN 200910093546 A CN200910093546 A CN 200910093546A CN 200910093546 A CN200910093546 A CN 200910093546A CN 101673339 A CN101673339 A CN 101673339A
- Authority
- CN
- China
- Prior art keywords
- spectrum
- absorption
- spectral
- rearrangement
- target
- 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
- 230000003595 spectral effect Effects 0.000 title claims abstract description 64
- 238000010521 absorption reaction Methods 0.000 title claims abstract description 63
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000004611 spectroscopical analysis Methods 0.000 title abstract 4
- 238000001228 spectrum Methods 0.000 claims abstract description 64
- 230000008707 rearrangement Effects 0.000 claims abstract description 39
- 238000000605 extraction Methods 0.000 claims abstract description 7
- 230000009466 transformation Effects 0.000 claims description 17
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 description 22
- 238000002310 reflectometry Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 5
- 241001122767 Theaceae Species 0.000 description 2
- 238000005452 bending Methods 0.000 description 2
- 230000008030 elimination Effects 0.000 description 2
- 238000003379 elimination reaction Methods 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000002087 whitening effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Images
Landscapes
- Investigating Or Analysing Materials By Optical Means (AREA)
- Spectrometry And Color Measurement (AREA)
Abstract
A target identification method of high spectroscopic data rearranged based on spectral absorption characteristics comprises the following steps: (1) reading the high spectroscopic data; (2) convertingminimal noise component; (3) removing spectrum continuously; (4) calculating absorption position, absorption depth and absorption characteristic left and right shoulders of pixel spectrum; (5) arranging from strong to weak according to the absorption depth by using the target spectrum to be identified as the base spectrum, and arranging the waveband without remarkable absorption characteristic from small to large according to wavelength to obtain the rearranged spectrum; (6) rearranging the spectroscopic characteristics of the high spectroscopic data according to the wavelength order of the target rearranged spectrum to be identified in the step (5) to obtain the characteristic extraction result based on the spectrum rearrangement; and (7) matching spectroscopic characteristic to obtain the target identification result. The method solves the problems of the influence on identification result caused by the factors of serious noisy influence on the traditional target identification method based on spectroscopic characteristic matching, unstable single characteristics and the like, thereby realizing reliable and accurate target identification.
Description
Technical Field
The invention relates to a hyperspectral data target identification method based on spectral absorption characteristic rearrangement, belongs to the technical field of hyperspectral data processing methods and identification application, and is suitable for theoretical methods and application technical research of hyperspectral data characteristic extraction and target identification.
Background
The physical and chemical properties of different ground objects are different, and the spectral characteristics are different, so that the fine spectral characteristics provided by the hyperspectral remote sensing data can be directly used as the basis for target identification. However, due to the influences of factors such as sunlight conditions, atmosphere, noise, instrument response and the like, the spectral characteristics can drift and become shallow, so that the spectral characteristics have great uncertainty, and how to apply hyperspectral data to realize stable and reliable direct identification of ground object types is one of the key problems to be solved in hyperspectral application.
In order to solve the problems, scholars at home and abroad develop a great deal of research in the field, at present, a hyperspectral remote sensing data target identification method is mainly based on a spectral feature matching method, and at present, the target identification method based on the spectral feature matching mainly comprises two categories: a target identification method based on parametric spectral feature matching and a target identification method based on spectral waveform matching. The target identification method based on spectral waveform matching cannot investigate the spectral difference of some local wave bands, so that the target identification method is greatly limited when identifying targets with similar spectral characteristics; the method based on parametric spectral feature matching mainly utilizes a single parametric feature as a recognition criterion, such as an absorption position or an absorption depth, and only utilizes a remarkable absorption feature, although the existing parametric spectral feature matching method based on parametric spectrum can effectively obtain local features of a ground feature spectrum, a single absorption feature is not stable enough, different types of ground features may have similar spectral features, and the spectral parametric spectrum feature is sensitive to noise, so that the existing target recognition method is difficult to meet the requirements of the existing hyperspectral application.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a hyperspectral data target identification method based on spectral absorption characteristic rearrangement, which is less affected by noise and comprehensively utilizes all absorption characteristics.
The invention relates to a hyperspectral data target identification method based on spectral absorption characteristic rearrangement, which adopts the technical scheme that: the method mainly utilizes minimum noise component transformation to reduce the interference of noise on spectral absorption characteristics, achieves spectral rearrangement by calculating the absorption position, the absorption depth, the left shoulder, the right shoulder and other absorption characteristics of hyperspectral data and sequencing from strong to weak according to the absorption depth, extracts the spectral characteristics, and finally utilizes a spectral matching method to achieve target identification based on the spectral rearrangement. The minimum noise component transformation effectively eliminates data noise and the correlation between wave bands; the spectrum rearrangement utilizes all absorption positions and depths, can effectively reflect local spectrum characteristic differences, and solves the problem of single characteristic instability.
The invention relates to a hyperspectral data target identification method based on spectral absorption characteristic rearrangement, which comprises the following steps:
(1) reading hyperspectral data;
(2) minimum noise component transformation;
(3) removing a spectrum continuum;
(4) calculating the absorption position, the absorption depth and the left and right shoulders of the absorption characteristics of the pixel spectrum;
(5) performing spectrum rearrangement according to absorption characteristics by taking the target spectrum to be identified as a base spectrum;
(6) carrying out spectral feature rearrangement on the hyperspectral data according to the wavelength sequence of the target rearrangement spectrum to be identified in the step (5) to obtain a feature extraction result based on the spectral rearrangement;
(7) and matching the spectral characteristics to obtain a target identification result.
Reading hyperspectral data in the step (1) comprises the following steps: x ═ X1,x2,…,xn]TAnd n is the number of pixels.
Wherein, the minimum noise component transformation in the step (2): the method is based on an estimation matrix of noise covariance, supposing that hyperspectral data read in the step (1) can be decomposed into X ═ Z + N, the matrix Z, N is an ideal signal matrix and a noise matrix respectively, and the transformation process is to adjust the value of noise and remove the correlation among wave bands; least noise component forward transform matrix ofWherein,ΛN (p)、UN (p)eigenvalue matrix, eigenvector matrix, U, of the noise component covariance matrix, respectivelyw (p)Is an eigenvector matrix of the covariance matrix after X whitening, and each wave band of the changed data T is arranged from large to small according to the SNR of the signal to noise ratioThe variance of the noise is 1, and no correlation exists between wave bands; the minimum noise component inverse transformation process principle is the same, and only the hyperspectral data T in the minimum noise component space is transformed to the original observation data space to obtain the data Y after the inverse transformation.
Wherein the spectral continuum described in step (3) removes: is a convex hull fit at the top of the spectrum, connecting the local spectral maxima with straight line segments. The first and last spectral data values are on the envelope, so the first and last bands in the continuum elimination curve of the output spectral continuum are equal to 1.0; the hyperspectral images after the continuum removal highlight the characteristic information of the spectrum of the ground object, eliminate the influence of different amplitudes of the spectrum of the same ground object caused by factors such as illumination conditions and the like, effectively inhibit noise and facilitate the calculation and matching of the spectral characteristics of the images.
The method for calculating the left shoulder and the right shoulder of the absorption position, the absorption depth and the absorption characteristic of the pixel spectrum in the step (4) comprises the following steps: by calculating the first order differential value, the spectrum bending point, the maximum and minimum spectrum reflectivity and the wavelength position thereof can be rapidly determined, the wavelength position of the maximum spectrum reflectivity is the left shoulder and the right shoulder of the absorption characteristic, the wavelength position of the minimum spectrum reflectivity is the absorption position, and the absorption depth d is 1-RmWherein R ismThe post spectral amplitude is removed for the continuum corresponding to the absorption position.
The spectrum rearrangement process based on the absorption characteristics by taking the target spectrum to be identified as the base spectrum in the step (5) is as follows: and (4) taking the target spectrum to be identified as a base spectrum, and arranging the absorption depths from strong to weak according to the absorption depth calculated in the step (4), wherein if no absorption characteristic exists, the absorption depths are arranged from small to large according to the wavelength.
And (3) performing spectral feature rearrangement on the hyperspectral data according to the wavelength sequence of the target rearrangement spectrum to be identified in the step (5) to obtain a feature extraction result based on the spectral rearrangement: and performing spectrum rearrangement on all the hyperspectral data according to the same sequence by taking the target spectrum rearrangement result to be identified as a reference.
The spectral feature matching in the step (7) obtains a target identification result: calculating the spectrum similarity after the spectrum rearrangement by adopting Orthogonal Projection Divergence (OPD), thereby realizing target identification; two P-dimensional spectral signals xi=[xi1,xi2,...,xiP]T,xj=[xj1,xj2,...,xjP]TThen P-dimensional spectral signal xiAnd xjThe orthogonal projection divergence between is expressed as:
wherein,
<math>
<mrow>
<msubsup>
<mi>P</mi>
<msub>
<mi>x</mi>
<mi>k</mi>
</msub>
<mo>⊥</mo>
</msubsup>
<mo>=</mo>
<msub>
<mi>I</mi>
<mrow>
<mi>P</mi>
<mo>×</mo>
<mi>P</mi>
</mrow>
</msub>
<mo>-</mo>
<msub>
<mi>x</mi>
<mi>k</mi>
</msub>
<msup>
<mrow>
<mo>(</mo>
<msubsup>
<mi>x</mi>
<mi>k</mi>
<mi>T</mi>
</msubsup>
<msub>
<mi>x</mi>
<mi>k</mi>
</msub>
<mo>)</mo>
</mrow>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<msubsup>
<mi>x</mi>
<mi>k</mi>
<mi>T</mi>
</msubsup>
<mo>,</mo>
</mrow>
</math>
k is I, j, and IP×PIs an identity matrix in dimension P < P >.
Compared with the prior art, the invention has the advantages that: the method overcomes the limitations that the traditional hyperspectral data target identification method is greatly influenced by noise, the matching identification characteristics are unstable and the like, and realizes high-reliability target identification by utilizing minimum noise component transformation and spectrum rearrangement. It has the following advantages: (1) minimum noise component transformation is introduced, and the problem that noise caused by factors such as atmosphere and remote sensors influences spectral feature extraction is solved; (2) before the characteristics are extracted, a continuum is used for removing, so that the influence of amplitude change caused by the problems of sunshine conditions, remote sensor response and the like is eliminated; (3) the target is identified by adopting spectral rearrangement and utilizing all absorption characteristics, so that the difficulty of accurate identification of the target caused by the problems of instability, ambiguity and the like of single characteristics is overcome.
Drawings
FIG. 1 is a flow chart of an implementation of a hyperspectral data target identification method based on spectral absorption characteristic rearrangement.
Detailed Description
In order to better explain the multi-class supervision and classification method of the hyperspectral remote sensing data, PHI aviation hyperspectral data is used for identifying the types of crops in Jiangsu foot tea field areas. The invention relates to a hyperspectral data target identification method based on spectral absorption characteristic rearrangement, which comprises the following concrete implementation steps of:
(1) reading hyperspectral data: reading PHI hyperspectral data of a square foot tea field, wherein the data size is 210 multiplied by 150 multiplied by 64, and the wave band interval is 455-805 nm;
(2) minimum noise component transformation: the method is based on an estimation matrix of noise covariance, supposing that hyperspectral data read in the step (1) can be decomposed into X ═ Z + N, the matrix Z, N is an ideal signal matrix and a noise matrix respectively, and the transformation process is to adjust the value of noise and remove the correlation among wave bands; least noise component forward transform matrix ofWherein,ΛN (64)、UN (64)eigenvalue matrix, eigenvector matrix, U, of the noise component covariance matrix, respectivelyw (64)Is the eigenvector matrix of the covariance matrix after X whitening, and each waveband of the changed data TArranged from large to small according to SNRThe variance of the noise is 1, and no correlation exists between wave bands; the minimum noise component inverse transformation process principle is the same, and only the hyperspectral data T in the minimum noise component space is transformed to the original observation data space to obtain the data Y after the inverse transformation.
(3) Spectral continuum removal: is a convex hull fit at the top of the spectrum, connecting the local spectral maxima with straight line segments. The first and last spectral data values are on the envelope, so the first and last bands in the continuum elimination curve of the output spectral continuum are equal to 1.0; the hyperspectral images after the continuum removal highlight the characteristic information of the spectrum of the ground object, eliminate the influence of different amplitudes of the spectrum of the same ground object caused by factors such as illumination conditions and the like, effectively inhibit noise and facilitate the calculation and matching of the spectral characteristics of the images.
(4) Calculating the absorption position, the absorption depth and the absorption characteristic left and right shoulders of the pixel spectrum: by calculating the first order differential value, the spectrum bending point, the maximum and minimum spectrum reflectivity and the wavelength position thereof can be rapidly determined, the wavelength position of the maximum spectrum reflectivity is the left shoulder and the right shoulder of the absorption characteristic, the wavelength position of the minimum spectrum reflectivity is the absorption position, and the absorption depth d is 1-RmWherein R ismThe post spectral amplitude is removed for the continuum corresponding to the absorption position.
(5) Performing spectrum rearrangement according to absorption characteristics by taking the target spectrum to be identified as a base spectrum: and (4) taking the target spectrum to be identified as a base spectrum, arranging according to the calculated absorption depth from strong to weak, and if no absorption characteristic exists, arranging according to the wavelength from small to large.
(6) And (5) carrying out spectral feature rearrangement on the hyperspectral data according to the wavelength sequence of the target rearrangement spectrum to be identified in the step (5) to obtain a feature extraction result based on the spectral rearrangement: and performing spectrum rearrangement on all the hyperspectral data according to the same sequence by taking the target spectrum rearrangement result to be identified as a reference.
(7) Matching spectral features to obtain a target identification result: calculating the spectrum similarity after the spectrum rearrangement by adopting an orthogonal projection divergence OPD (optical phase detector), thereby realizing target identification; two P-dimensional spectral signals xi=[xi1,xi2,...,xi64]T,xj=[xj1,xj2,...,xj64]TAnd then 64-dimensional spectral signal xiAnd xjThe orthogonal projection divergence between is expressed as:
wherein,
<math>
<mrow>
<msubsup>
<mi>P</mi>
<msub>
<mi>x</mi>
<mi>k</mi>
</msub>
<mo>⊥</mo>
</msubsup>
<mo>=</mo>
<msub>
<mi>I</mi>
<mrow>
<mn>64</mn>
<mo>×</mo>
<mn>64</mn>
</mrow>
</msub>
<mo>-</mo>
<msub>
<mi>x</mi>
<mi>k</mi>
</msub>
<msup>
<mrow>
<mo>(</mo>
<msubsup>
<mi>x</mi>
<mi>k</mi>
<mi>T</mi>
</msubsup>
<msub>
<mi>x</mi>
<mi>k</mi>
</msub>
<mo>)</mo>
</mrow>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<msubsup>
<mi>x</mi>
<mi>k</mi>
<mi>T</mi>
</msubsup>
<mo>,</mo>
</mrow>
</math>
k is I, j, and I64×64Is an identity matrix of 64 x 64 dimensions.
Claims (3)
1. A hyperspectral data target identification method based on spectral absorption characteristic rearrangement is characterized by comprising the following steps: it comprises the following steps:
(1) reading hyperspectral data;
(2) minimum noise component transformation;
(3) removing a spectrum continuum;
(4) calculating the absorption position, the absorption depth and the left and right shoulders of the absorption characteristics of the pixel spectrum;
(5) performing spectrum rearrangement according to absorption characteristics by taking the target spectrum to be identified as a base spectrum;
(6) carrying out spectral feature rearrangement on the hyperspectral data according to the wavelength sequence of the target rearrangement spectrum to be identified in the step (5) to obtain a feature extraction result based on the spectral rearrangement;
(7) and matching the spectral characteristics to obtain a target identification result.
2. The hyperspectral data object identification method based on spectral absorption feature rearrangement of claim 1 wherein the minimum noise component transformation in step (2) is first a minimum noise component forward transformation and then a minimum noise component inverse transformation on the transformed data in order to eliminate the influence of noise on the spectral absorption feature calculation.
3. The hyperspectral data object identification method based on spectral absorption feature rearrangement according to claim 1, wherein the method for performing spectral rearrangement according to the absorption features in the step (5) comprises the following steps: and sequencing the target spectrum to be identified from strong to weak according to the absorption depth by taking the target spectrum to be identified as a base spectrum, and sequencing the characteristic wave bands without obvious absorption from small to large according to the wavelength.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN200910093546A CN101673339A (en) | 2009-09-25 | 2009-09-25 | Target identification method of high spectroscopic data rearranged based on spectral absorption characteristics |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN200910093546A CN101673339A (en) | 2009-09-25 | 2009-09-25 | Target identification method of high spectroscopic data rearranged based on spectral absorption characteristics |
Publications (1)
Publication Number | Publication Date |
---|---|
CN101673339A true CN101673339A (en) | 2010-03-17 |
Family
ID=42020561
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN200910093546A Pending CN101673339A (en) | 2009-09-25 | 2009-09-25 | Target identification method of high spectroscopic data rearranged based on spectral absorption characteristics |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN101673339A (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101916377A (en) * | 2010-07-15 | 2010-12-15 | 北京航空航天大学 | Method for intelligently identifying mineral information based on spectral information |
CN102385694A (en) * | 2010-09-06 | 2012-03-21 | 邬明权 | Hyperspectral identification method for land parcel-based crop variety |
CN102521830A (en) * | 2011-11-30 | 2012-06-27 | 浙江大学 | Optimum band selection method for hyperspectral images of canopy of crop under disease stress |
CN102609711A (en) * | 2012-02-21 | 2012-07-25 | 核工业北京地质研究院 | Information extraction method applicable to hyperspectral image |
CN102682430A (en) * | 2012-04-25 | 2012-09-19 | 北京航空航天大学 | Non-uniform correction method combining spectrum and spatial information for high-spectral data |
CN109580497A (en) * | 2018-12-13 | 2019-04-05 | 中国国土资源航空物探遥感中心 | A kind of high spectrum mineral geochemical anomalies studying method based on singular integrations |
CN110378268A (en) * | 2019-07-10 | 2019-10-25 | 中国科学院长春光学精密机械与物理研究所 | EO-1 hyperion fast target detection method and system |
CN114371138A (en) * | 2022-01-13 | 2022-04-19 | 中国自然资源航空物探遥感中心 | Hyperspectral remote sensing spectral characteristic absorption position determining method and system |
-
2009
- 2009-09-25 CN CN200910093546A patent/CN101673339A/en active Pending
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101916377B (en) * | 2010-07-15 | 2012-12-05 | 北京航空航天大学 | Method for intelligently identifying mineral information based on spectral information |
CN101916377A (en) * | 2010-07-15 | 2010-12-15 | 北京航空航天大学 | Method for intelligently identifying mineral information based on spectral information |
CN102385694A (en) * | 2010-09-06 | 2012-03-21 | 邬明权 | Hyperspectral identification method for land parcel-based crop variety |
CN102521830A (en) * | 2011-11-30 | 2012-06-27 | 浙江大学 | Optimum band selection method for hyperspectral images of canopy of crop under disease stress |
CN102521830B (en) * | 2011-11-30 | 2013-11-06 | 浙江大学 | Optimum band selection method for hyperspectral images of canopy of crop under disease stress |
CN102609711A (en) * | 2012-02-21 | 2012-07-25 | 核工业北京地质研究院 | Information extraction method applicable to hyperspectral image |
CN102682430A (en) * | 2012-04-25 | 2012-09-19 | 北京航空航天大学 | Non-uniform correction method combining spectrum and spatial information for high-spectral data |
CN102682430B (en) * | 2012-04-25 | 2014-07-09 | 北京航空航天大学 | Non-uniform correction method combining spectrum and spatial information for high-spectral data |
CN109580497A (en) * | 2018-12-13 | 2019-04-05 | 中国国土资源航空物探遥感中心 | A kind of high spectrum mineral geochemical anomalies studying method based on singular integrations |
CN109580497B (en) * | 2018-12-13 | 2020-01-03 | 中国自然资源航空物探遥感中心 | Hyperspectral mineral abnormal information extraction method based on singularity theory |
CN110378268A (en) * | 2019-07-10 | 2019-10-25 | 中国科学院长春光学精密机械与物理研究所 | EO-1 hyperion fast target detection method and system |
CN110378268B (en) * | 2019-07-10 | 2022-04-26 | 中国科学院长春光学精密机械与物理研究所 | Hyperspectral rapid target detection method and system |
CN114371138A (en) * | 2022-01-13 | 2022-04-19 | 中国自然资源航空物探遥感中心 | Hyperspectral remote sensing spectral characteristic absorption position determining method and system |
CN114371138B (en) * | 2022-01-13 | 2024-04-12 | 中国自然资源航空物探遥感中心 | Hyperspectral remote sensing spectral characteristic absorption position determining method and hyperspectral remote sensing spectral characteristic absorption position determining system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN101673339A (en) | Target identification method of high spectroscopic data rearranged based on spectral absorption characteristics | |
CN106019214B (en) | Wide-band coherent signal source DOA estimation method | |
CN103456015B (en) | A kind of SAR target detection method based on optimum score field Gabor spectrum signature | |
CN109490223B (en) | Target detection and identification system and method based on programmable hyperspectral imaging | |
CN104182978B (en) | A kind of high spectrum image object detection method based on empty spectrum nuclear sparse expression | |
CN107038436B (en) | Hyperspectral image target detection method based on tensor spectrum matched filtering | |
CN107194936B (en) | Hyperspectral image target detection method based on superpixel combined sparse representation | |
CN101252382A (en) | Wide frequency range signal polarizing and DOA estimating method and apparatus | |
CN115144357A (en) | Mineral identification method based on airborne mid-infrared hyperspectral remote sensing data of mining area | |
Feng et al. | A Coarse-to-fine Hyperspectral Target Detection Method based on Low Rank Tensor Decomposition | |
CN103926203A (en) | Spectral angle mapping method aiming at ground object spectrum uncertainty | |
CN103268476B (en) | A kind of Remote Sensing Target monitoring method | |
CN102880861B (en) | High-spectrum image classification method based on linear prediction cepstrum coefficient | |
Yumus et al. | Land cover classification for synthetic aperture radar imagery by using unsupervised methods | |
CN101515330B (en) | Discrimination-oriented correlated characters fusion recognition system of an image | |
CN104408476A (en) | Deep sparse main component analysis-based polarimetric SAR image classification method | |
CN105427351A (en) | High spectral image compression sensing method based on manifold structuring sparse prior | |
Ghosh et al. | Fractal-based dimensionality reduction of hyperspectral images | |
CN104732190A (en) | Synthetic aperture sonar target detection method based on orthogonal texture correlation analysis | |
CN105303009A (en) | Super-resolution spectrum estimation method based on compressed sensing and regular MFOCUSS | |
CN109584255B (en) | Target contour extraction method based on fusion of panchromatic image and hyperspectral image | |
Zhao et al. | Endmember-specified virtual dimensionality in hyperspectral imagery | |
Geng et al. | Principle of small target detection for hyperspectral imagery | |
Zhou et al. | Research on modified SVM for classification of SAR images | |
Ren et al. | Feature extraction with modified Fisher’s linear discriminant analysis |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
WD01 | Invention patent application deemed withdrawn after publication |
Open date: 20100317 |