高光谱遥感岩矿特征提取与分类方法研究
|
摘要
高光谱遥感图像包含了丰富的空间、辐射和光谱三重信息，光谱分辨率在λ/100数量级，在电磁波谱的紫外、可见光、近红外和中红外区域，获取许多非常窄且光谱连续的图像数据。高光谱遥感图像波段众多、信息量大，对地物光谱特性的测度细致、对物质的描述精确，为地物识别带来了一定的优势。但是波段间的信息相关性强、信息冗余多，也带来了一些挑战。高光谱遥感图像分类中一个重要的现象就是维数灾难（Hughes现象）。当用传统的监督分类方法对高光谱图像进行分类时，随着波段数目的增加，需要训练样本的数量急剧增加，获得大量的训练样本在高光谱遥感图像中要花费很多人力物力，有时甚至是很难做到的。
高光谱遥感特征提取可以缓解Hughes现象，它一方面可以压缩数据量，并完成去噪的工作；另一方面通过特征提取，使目标的光谱特征更加明显，更有利于后续分类及信息提取工作。主成分分析（PCA）、最大/最小自相关因子法(MAF)、最小噪声分离变换（MNF）等线性变换法在高光谱遥感特征提取中应用广泛。但是高光谱遥感图像很多时候具有非线性的特点，线性特征提取方法容易照成信息丢失和失真。Scholkopf等于1998年对PCA进行扩展发展了非线性的核主成分分析（KPCA），此后国内外对KPCA展开了大量的研究，但是关于核最大/最小自相关因子法(KMAF)、核最小噪声分离变换（KMNF）的研究少见。
高光谱遥感是一个崭新的领域，其数据绝大多数都是收费的。本文侧重理论与方法的研究，选用资料丰富的美国内华达州Cuprite矿区免费的AVIRIS高光谱遥感岩矿图像作为原始数据，研究基于核方法（KPCA、KMAF、KMNF）的高光谱图像岩矿特征提取以压缩数据量、提取对后续分类有利的特征；在特征提取后的数据上进行端元提取；研究针对高光谱遥感图像的分类及信息提取方法，主要是基于波谱特征和支持向量机（SVM）的高光谱图像分类。
本论文取得的主要研究成果与创新认识如下：
（1）设计及实现了基于核方法（KPCA、KMAF、KMNF）的高光谱图像特征提取方法。在PCA、MAF、MNF算法的基础上，引入核方法，发展KPCA、KMAF、KMNF非线性特征提取算法，并进行高光谱图像特征提取实验，探讨其参数对高光谱图像特征提取的影响，对比PCA、KPCA方法，KPCA、KMAF/KMNF方法进行高光谱图像特征提取的效果。表明KPCA、KMAF、KMNF算法的参数σ对算法的运算时间，没有大的影响；随着样本数目的增加算法的运行时间增加较快，但是在小样本情况下合理设置参数，KPCA、KMAF、KMNF特征提取算法即可以取得较好的效果；PCA、KPCA方法降维速度快，但降维后波段不是严格按图像质量排序； KMAF、KMNF降维效率低于PCA、KPCA，但是严格按图像质量排序，对于波段选择及后续图像分类，信息提取都很有利。
（2）设计及实现了基于PCA、KPCA、KMAF/KMNF的PPI端元提取。基于不同的特征提取后的图像，采用PPI这种指导端元提取的方法，并利用N维散度分析人机交互最终确定端元并进行端元识别，发现：基于PCA特征提取后的图像由于压缩过快，可能丢失某些信息，在利用PPI进行端元提取的过程中会较多的依赖操作者的技巧，并且可能丢失部分端元；基于KPCA、KMAF/KMNF的端元提取考虑了图像的非线性特征，可以较多的提取端元； KMAF/KMNF变换后的图像严格按照图像质量排序，操作过程相对简单，且端元提取的效果较好。
（3）实现了基于波谱特征的高光谱图像分类。利用编码匹配、光谱角填图（SAM）、波谱特征拟合（SFF）、匹配滤波（MF）、混合调制匹配滤波（MTMF）方法进行了高光谱图像分类实验及精度评价。表明编码匹配只适用于粗略的分类和识别；匹配滤波是一种快速的分类方法，会产生较多的虚假信号；SFF是一种基于吸收特征的分类方法，对于吸收特征比较明显的矿物识别率较高；SAM夹角值与光谱向量的模无关，也就是与图像的增益系数无关，只比较光谱在形状上的相似性；MTMF是MF和线性混合理论的组合，减少了MF虚假信号出现的概率，具有较高的分类精度。
（4）对支持向量机（SVM）算法进行了改进，并将其应用到高光谱图像分类及信息提取中。深入探讨支持向量机的理论基础及其分类的基本原理。采用一对一方法将两类支持向量机问题推广到多类问题进行高光谱图像分类，并利用收缩和缓存技术来提高其效率。进行了基于SVM的高光谱图像分类实验，分析了数据维数、核函数及样本个数对分类的影响。发现SVM分类方法受数据的维度影响较小，具有一定的抗噪声能力；不同核函数对SVM分类结果影响不大；利用SVM进行分类时，合理设置有关参数，在小样本情况下也可取得较高精度，显示了利用SVM进行高光谱图像分类的优越性。
（5）初步形成了一套科学实用的利用高光谱遥感图像提取岩矿弱信息的方法与技术流程。即首先对高光谱遥感岩矿图像进行辐射较正得到反射率数据；然后用核方法进行特征提取，达到降维与突出光谱特征的目的；基于降维后的数据提取端元并进行识别；最后基于波谱特征及SVM进行图像分类及信息提取。
Space, radiation and spectral information of hyperspectral remote sensing imagesare rich, whose spectral resolution is at λ/100orders of magnitude. A number of verynarrow and contiguous spectral image data can obtained in the electromagneticspectrum of ultraviolet, visible, near infrared and mid-infrared region. Hyperspectralremote sensing images have lots of bands, can describe surface features’ spectralproperties in detail which bring some advantages in surface features identification.But the bands are strongly correlated and information is redundant which also bringsome challenges. Dimension disaster(Hughes phenomenon) is usually occurred inhyperspectral remote sensing image classification.When traditional supervisedclassification methods are used for hyperspectral image classification, trainingsamples is increased dramatically with the increase of band number. Lots oftraining samples are difficult to obtained in hyperspectral remote sensing images.
Hyperspectral remote sensing feature extraction can alleviate Hughes phenomenon,it can compress data and remove noise of data on the one hand; on the other hand,spectral characteristics of the target is more obvious through feature extraction, it ismore conductive to the subsequent classification and information extraction. Linearfeature transformation methods such as principal component analysis(PCA), min/maxautocorrelation factors(MAF), minimum noise fraction(MNF) are used widely inhyperspectral remote sensing feature extraction. But they are likely to result indistortion and loss of data information for non-linear hyperspectral remote sensingimage data. Scholkopf expanded principal component analysis to nonlinear kernelprincipal component analysis (KPCA) in1998, from then on KPCA is studied homeand abroad, but few researchers studied kernel min/max autocorrelationfactors(KMAF) and kernel minimum noise fraction(KMNF).
Hyperspectral remote sensing is a new area, majority of the data are charged. Thedissertation focuses on theory and methods, so the free AVIRIS hyperspectral remotesensing images in Cuprite, Nevada, United States are used as data source. The paperstudies hyperspectral mineral feature extraction based on kernel methods (KPCA,KMAF, KMNF); extracts end-member after feature extraction; studies hyperspectral image classification and information extraction based on spectral features and supportvector machine (SVM).
The dissertation has the following achievements and innovations:
(1) Design and implementation of hyperspectral image feature extraction based onkernel methods (KPCA, KMAF, KMNF). We introduced kernel methods in PCA,MAF, MNF algorithms, developed KPCA, KMAF, KMNF nonlinear featureextraction algorithms and did experiments to study their parameters, compared PCAand KPCA, KPCA and KMAF/KMNF. It showes that KPCA, KMAF and KMNFparameterσhas little influence on algorithms’ time efficency; with the increase ofsample number running time increases rapidly, but hyperspectral image featureextraction based on KPCA、KMAF、KMNF can get good results with small samples;PCA, KPCA methods reduct data dimension quickly, but images after it are notstrictly sorted by their quality; KMAF, KMNF reduct data slower than PCA,KPCA,but images after it are strictly sorted by their quality, this is very usefull for bandselection, image classification and information extraction.
(2) Design and implementation of PPI endmember extraction based on PCA,KPCA, KMAF/KMNF. Based on different images after feature extraction, use PPIand N-dimensional visualizer to extract endmembers. We find that PPI endmemberextraction based on PCA depends more on operators’ skills and may lost someendmembers because PCA compress data quickly and may lost image information;PPI endmember extraction based on KPCA,KMAF/KMNF can get more endmembersbecause they consider image non-linear characteristics; PPI endmember extractionbased on KMAF/KMNF are easy and efficent because images are strictly sorted bytheir quality.
(3)Implementation of hyperspectral image classification based on spectralcharacteristics. Binary encoding, spectral angle mapper (SAM), matchedfiltering(MF), mixture-tuned matched filtering(MTMF）were used in hyperspectralimage classification, accuracy assessments were done. We find that binary encoding issuitable for hyperspectral image roughly classification; MF is a fast classificationmethod and can produce more false signals; SFF is an absorption feature basedclassification method, it can identify minerals which have obvious absorption featureefficently; SAM angle values have nothing to do with spectrum vector modes, itsimply compares the spectrum shape similarity; MTMF is the combination of MF andlinear mixture theory, it reduces the false signal appeared in MF and can get highclassification accuracy.
(4) Hyperspectral image classification and information extraction based onimproved support vector machine (SVM). The theoretical basis of support vectormachines and its classification principles are studied. One-against-one method is usedto extend the two-class support vector machine to multi-class in hyperspectral imageclassification, shringking and caching techniques are used to improve its efficiency. SVM-based hyperspectral image classification experiments are done and influence ofdata dimension, kenel function and sample number are analyzed. We find that datadimension has little influence on SVM-based classification, and it has some noiseimmunity; different kernel functions may get similar classification accuracy. Withreasonable parameters, we can get high classification accuracy when sample numberis small. All of this shows superiority of SVM-based hyperspectral imageclassification.
（5）Initially forms a set of scientific and practical mineral weak informationextraction methods and technology based on hyperspectral remote sensing images.Fist get hyperspectral remote sensing reflectance data by radiometric correction; thenuse kernel methods for feature extraction, so data was reducted; extract endmemberbased on reducted data; finally based on spectral characteristics and SVM for imageclassification and information extraction.
高光谱遥感特征提取可以缓解Hughes现象，它一方面可以压缩数据量，并完成去噪的工作；另一方面通过特征提取，使目标的光谱特征更加明显，更有利于后续分类及信息提取工作。主成分分析（PCA）、最大/最小自相关因子法(MAF)、最小噪声分离变换（MNF）等线性变换法在高光谱遥感特征提取中应用广泛。但是高光谱遥感图像很多时候具有非线性的特点，线性特征提取方法容易照成信息丢失和失真。Scholkopf等于1998年对PCA进行扩展发展了非线性的核主成分分析（KPCA），此后国内外对KPCA展开了大量的研究，但是关于核最大/最小自相关因子法(KMAF)、核最小噪声分离变换（KMNF）的研究少见。
高光谱遥感是一个崭新的领域，其数据绝大多数都是收费的。本文侧重理论与方法的研究，选用资料丰富的美国内华达州Cuprite矿区免费的AVIRIS高光谱遥感岩矿图像作为原始数据，研究基于核方法（KPCA、KMAF、KMNF）的高光谱图像岩矿特征提取以压缩数据量、提取对后续分类有利的特征；在特征提取后的数据上进行端元提取；研究针对高光谱遥感图像的分类及信息提取方法，主要是基于波谱特征和支持向量机（SVM）的高光谱图像分类。
本论文取得的主要研究成果与创新认识如下：
（1）设计及实现了基于核方法（KPCA、KMAF、KMNF）的高光谱图像特征提取方法。在PCA、MAF、MNF算法的基础上，引入核方法，发展KPCA、KMAF、KMNF非线性特征提取算法，并进行高光谱图像特征提取实验，探讨其参数对高光谱图像特征提取的影响，对比PCA、KPCA方法，KPCA、KMAF/KMNF方法进行高光谱图像特征提取的效果。表明KPCA、KMAF、KMNF算法的参数σ对算法的运算时间，没有大的影响；随着样本数目的增加算法的运行时间增加较快，但是在小样本情况下合理设置参数，KPCA、KMAF、KMNF特征提取算法即可以取得较好的效果；PCA、KPCA方法降维速度快，但降维后波段不是严格按图像质量排序； KMAF、KMNF降维效率低于PCA、KPCA，但是严格按图像质量排序，对于波段选择及后续图像分类，信息提取都很有利。
（2）设计及实现了基于PCA、KPCA、KMAF/KMNF的PPI端元提取。基于不同的特征提取后的图像，采用PPI这种指导端元提取的方法，并利用N维散度分析人机交互最终确定端元并进行端元识别，发现：基于PCA特征提取后的图像由于压缩过快，可能丢失某些信息，在利用PPI进行端元提取的过程中会较多的依赖操作者的技巧，并且可能丢失部分端元；基于KPCA、KMAF/KMNF的端元提取考虑了图像的非线性特征，可以较多的提取端元； KMAF/KMNF变换后的图像严格按照图像质量排序，操作过程相对简单，且端元提取的效果较好。
（3）实现了基于波谱特征的高光谱图像分类。利用编码匹配、光谱角填图（SAM）、波谱特征拟合（SFF）、匹配滤波（MF）、混合调制匹配滤波（MTMF）方法进行了高光谱图像分类实验及精度评价。表明编码匹配只适用于粗略的分类和识别；匹配滤波是一种快速的分类方法，会产生较多的虚假信号；SFF是一种基于吸收特征的分类方法，对于吸收特征比较明显的矿物识别率较高；SAM夹角值与光谱向量的模无关，也就是与图像的增益系数无关，只比较光谱在形状上的相似性；MTMF是MF和线性混合理论的组合，减少了MF虚假信号出现的概率，具有较高的分类精度。
（4）对支持向量机（SVM）算法进行了改进，并将其应用到高光谱图像分类及信息提取中。深入探讨支持向量机的理论基础及其分类的基本原理。采用一对一方法将两类支持向量机问题推广到多类问题进行高光谱图像分类，并利用收缩和缓存技术来提高其效率。进行了基于SVM的高光谱图像分类实验，分析了数据维数、核函数及样本个数对分类的影响。发现SVM分类方法受数据的维度影响较小，具有一定的抗噪声能力；不同核函数对SVM分类结果影响不大；利用SVM进行分类时，合理设置有关参数，在小样本情况下也可取得较高精度，显示了利用SVM进行高光谱图像分类的优越性。
（5）初步形成了一套科学实用的利用高光谱遥感图像提取岩矿弱信息的方法与技术流程。即首先对高光谱遥感岩矿图像进行辐射较正得到反射率数据；然后用核方法进行特征提取，达到降维与突出光谱特征的目的；基于降维后的数据提取端元并进行识别；最后基于波谱特征及SVM进行图像分类及信息提取。
Space, radiation and spectral information of hyperspectral remote sensing imagesare rich, whose spectral resolution is at λ/100orders of magnitude. A number of verynarrow and contiguous spectral image data can obtained in the electromagneticspectrum of ultraviolet, visible, near infrared and mid-infrared region. Hyperspectralremote sensing images have lots of bands, can describe surface features’ spectralproperties in detail which bring some advantages in surface features identification.But the bands are strongly correlated and information is redundant which also bringsome challenges. Dimension disaster(Hughes phenomenon) is usually occurred inhyperspectral remote sensing image classification.When traditional supervisedclassification methods are used for hyperspectral image classification, trainingsamples is increased dramatically with the increase of band number. Lots oftraining samples are difficult to obtained in hyperspectral remote sensing images.
Hyperspectral remote sensing feature extraction can alleviate Hughes phenomenon,it can compress data and remove noise of data on the one hand; on the other hand,spectral characteristics of the target is more obvious through feature extraction, it ismore conductive to the subsequent classification and information extraction. Linearfeature transformation methods such as principal component analysis(PCA), min/maxautocorrelation factors(MAF), minimum noise fraction(MNF) are used widely inhyperspectral remote sensing feature extraction. But they are likely to result indistortion and loss of data information for non-linear hyperspectral remote sensingimage data. Scholkopf expanded principal component analysis to nonlinear kernelprincipal component analysis (KPCA) in1998, from then on KPCA is studied homeand abroad, but few researchers studied kernel min/max autocorrelationfactors(KMAF) and kernel minimum noise fraction(KMNF).
Hyperspectral remote sensing is a new area, majority of the data are charged. Thedissertation focuses on theory and methods, so the free AVIRIS hyperspectral remotesensing images in Cuprite, Nevada, United States are used as data source. The paperstudies hyperspectral mineral feature extraction based on kernel methods (KPCA,KMAF, KMNF); extracts end-member after feature extraction; studies hyperspectral image classification and information extraction based on spectral features and supportvector machine (SVM).
The dissertation has the following achievements and innovations:
(1) Design and implementation of hyperspectral image feature extraction based onkernel methods (KPCA, KMAF, KMNF). We introduced kernel methods in PCA,MAF, MNF algorithms, developed KPCA, KMAF, KMNF nonlinear featureextraction algorithms and did experiments to study their parameters, compared PCAand KPCA, KPCA and KMAF/KMNF. It showes that KPCA, KMAF and KMNFparameterσhas little influence on algorithms’ time efficency; with the increase ofsample number running time increases rapidly, but hyperspectral image featureextraction based on KPCA、KMAF、KMNF can get good results with small samples;PCA, KPCA methods reduct data dimension quickly, but images after it are notstrictly sorted by their quality; KMAF, KMNF reduct data slower than PCA,KPCA,but images after it are strictly sorted by their quality, this is very usefull for bandselection, image classification and information extraction.
(2) Design and implementation of PPI endmember extraction based on PCA,KPCA, KMAF/KMNF. Based on different images after feature extraction, use PPIand N-dimensional visualizer to extract endmembers. We find that PPI endmemberextraction based on PCA depends more on operators’ skills and may lost someendmembers because PCA compress data quickly and may lost image information;PPI endmember extraction based on KPCA,KMAF/KMNF can get more endmembersbecause they consider image non-linear characteristics; PPI endmember extractionbased on KMAF/KMNF are easy and efficent because images are strictly sorted bytheir quality.
(3)Implementation of hyperspectral image classification based on spectralcharacteristics. Binary encoding, spectral angle mapper (SAM), matchedfiltering(MF), mixture-tuned matched filtering(MTMF）were used in hyperspectralimage classification, accuracy assessments were done. We find that binary encoding issuitable for hyperspectral image roughly classification; MF is a fast classificationmethod and can produce more false signals; SFF is an absorption feature basedclassification method, it can identify minerals which have obvious absorption featureefficently; SAM angle values have nothing to do with spectrum vector modes, itsimply compares the spectrum shape similarity; MTMF is the combination of MF andlinear mixture theory, it reduces the false signal appeared in MF and can get highclassification accuracy.
(4) Hyperspectral image classification and information extraction based onimproved support vector machine (SVM). The theoretical basis of support vectormachines and its classification principles are studied. One-against-one method is usedto extend the two-class support vector machine to multi-class in hyperspectral imageclassification, shringking and caching techniques are used to improve its efficiency. SVM-based hyperspectral image classification experiments are done and influence ofdata dimension, kenel function and sample number are analyzed. We find that datadimension has little influence on SVM-based classification, and it has some noiseimmunity; different kernel functions may get similar classification accuracy. Withreasonable parameters, we can get high classification accuracy when sample numberis small. All of this shows superiority of SVM-based hyperspectral imageclassification.
（5）Initially forms a set of scientific and practical mineral weak informationextraction methods and technology based on hyperspectral remote sensing images.Fist get hyperspectral remote sensing reflectance data by radiometric correction; thenuse kernel methods for feature extraction, so data was reducted; extract endmemberbased on reducted data; finally based on spectral characteristics and SVM for imageclassification and information extraction.
引文
[1]陈伟,余旭初,王鹤,等.基于粒子群算法的高光谱影像端元提取技术[J],测绘科学,2011,36(4):16-19.
[2]测绘科技发展“十二五”规划（征求意见稿）.国家测绘地理信息局科技与国际合作司.2010年11月.
[3]杜培军,陈云浩,方涛,等.高光谱遥感数据光谱特征的提取与应用[J].中国矿业大学学报[J],2003,32(5):500-504.
[4]冯静,舒宁.改进型遗传算法和支持向量机的波段选择研究[J],武汉理工大学学报,2009,31(18):120-123.
[5]冯静,舒宁.基于新颖子波变换的高光谱遥感图像特征提取[J].中国图象图形学报,2009,14(10):2042-2046.
[6]甘甫平,王润生,马蔼乃.基于特征谱带的高光谱遥感矿物谱系识别[J].地学前缘(中国地质大学,北京),2003,10(2):445-454.
[7]官莉,刘旸,张雪慧.人工神经网络算法在红外高光谱资料反演大气温度廓线中的应用[J],大气科学学报,2010,33(3):341-346.
[8]耿修瑞，赵永超，刘素红,等.高维义积的矩阵计算以及在高光谱图像端元自动提取中的应用[J].中国科学:信息科学，2010，40:646-652
[9]李建民,张钹,林福宗.支持向量机的训练算法[J].清华大学学报(自然科学版),2003,43(1):120-124.
[10]李素,李文正,周建军,等.遥感影像混合像元分解中的端元选择方法综述[J].地理与地理信息科学,2007,23(5):35-39.
[11]李姗姗,田庆.高光谱遥感图像的端元递进提取算法[J],遥感学报，2009,13(2)：269-275
[12]李庆震,祝小平.基于核PCA的智能图像分析算法[J].弹箭与制导学报,2007,27(5):189-193
[13]刘颖,谷延锋,张晔.基于改进遗传算法的超光谱图像特征选择方法[J].哈尔滨工业大学学报,2005,37(6):733-736.
[14]刘卓,易东云.投影寻踪方法与高光谱遥感图像数据特征提取的研究[J].数学理论与应用,2003,23(1):76-81.
[15]刘正军,王长耀,张继贤.基于小波与遗传算法的特征提取与特征选择[J].遥感学报,2005,9(2):176-185.
[16]刘正军.高维遥感数据土地覆盖特征提取与分类研究[D].北京:中国科学院遥感应用研究所，2003.
[17]刘汉湖.岩矿波谱数据分析与信息提取方法研究[D].成都:成都理工大学,2008.
[18]刘翔.基于光谱维变换的高光谱图像目标探测研究[D].北京:中国科学院研究生院,2008.
[19]刘江永,王大明.基于支持向量机的快速高光谱分类研究[J],陕西师范大学学报(自然科学版),2009,37(4):43-47.
[20]骆剑承,周成虎,梁怡,等.支撑向量机及其遥感影像空间特征提取和分类的应用研究[J].遥感学报,2002,6(1):50-55.
[21]梁顺林.定量遥感[M],北京:科学出版社,2009.
[22]路威.面向目标探测的高光谱影像特征提取与分类技术研究[D].郑州:解放军信息工程大学,2005.
[23]林娜,杨武年,刘汉湖.基于高光谱遥感的岩矿端元识别及信息提取研究[J].遥感信息,2011,5:114-118.
[24]牟少敏.核方法的研究及其应用[D].北京：北方交通大学，2008.
[25]梅安新，彭望琭，秦其明,等.遥感导论[M].北京:高等教育出版社,2001.
[26]牛鹏,魏维.基SVM动态集成的高光谱遥感图像分类[J].计算机应用,2010,30(6):1590-1593.
[27]佘红伟，张艳宁，袁和金.一种无监督高光谱图像分类算法[J].中国图像图形学报,2008,13(6):1123-1128.
[28]苏红军,杜培军.高光谱数据特征选择与特征提取研究[J].遥感技术与应用，2006,21(4):288-293.
[29]沈照庆,陶建斌.基于模糊核主成分分析的高光谱遥感影像特征提取研究[J].国土资源遥感,2009,(3):41-45.
[30]谭琨,杜培军.基于再生核Hilbert空间小波核函数支持向量机的高光谱遥感影像分类[J].测绘学报,2011,40(2):142-147.
[31]童庆禧,张兵,郑兰芬.高光谱遥感:原理、技术与应用[M].北京:高等教育出版社,2006.
[32]田庆久,宫鹏.地物波谱数据库研究现状与发展趋势[J].遥感信息，2002，3：1-6.
[33]王润生,甘甫平,闫柏琨.高光谱矿物填图技术与应用研究[J].国土资源遥感,2010,83(1):1-13.
[34]万家强,王越,刘羽.改进KPCA对分类数据的特征提取[J],计算机工程与设计，2010，31（18）：4085-4088.
[35]王国胜.核函数的性质及其构造方法[J].计算机科学，2006，33(6):172-178.
[36]王毅,张良培,李平湘.基于自动搜索和光谱匹配技术的训练样本纯化算法[J].武汉大学学报(信息科学版),2007(3):216-219.
[37]吴波,张良培,李平湘.基于支撑向量回归的高光谱混合像元非线性分解[J].遥感学报,2006,10(3):312-318.
[38]吴波,张良培,李平湘.高光谱端元自动提取的迭代分解方法[J],遥感学报,2005,(3).
[39]熊桢,童庆禧,郑兰芬.用于高光谱遥感图像分类的一种高阶神经网络算法[J].中国图象图形学报,2000,(3)196-201.
[40]徐宏根,马洪超,李德仁.结合SOM神经网络和混合像元分解的高光谱影像分类方法研究[J].遥感学报,2007,11(6):778-786.
[41]徐芳.航空影像分割的支持向量机方法[D].武汉:武汉大学,2004.
[42]徐元进.面向找矿的高光谱遥感岩矿信息提取方法研究[D].武汉:中国地质大学,2009.
[43]薛宁静.多类支持向量机分类器对比研究[J].计算机工程与设计,2011,32(5):1792-1795
[44]闰永忠,万余庆.高光谱图像模糊识别分类及其精度评价[J].地球信息科学,2005,7(4):20-24.
[45]杨诸胜,郭雷,罗欣,等.一种基于主成分分析的高光谱图像波段选择算法[J].微电子学与计算机,2006,23(12):72-74.
[46]杨飞,张柏,刘志明,等.玉米冠层FPAR的高光谱遥感估算研究——基于PCA方法及近、短波红外波段[J].国土资源遥感,2008,(4):9-13.
[47]杨国鹏.基于核方法的高光谱影像分类与特征提取[D],郑州：解放军信息工程大学，2007.
[48]杨国鹏.基于机器学习方法的高光谱影像分类研究[D].河南:解放军信息工程大学,2010.
[49]尹峻松,肖健,周宗潭,等.非线性流形学习方法的分析与应用[J],自然科学进展,2007,17(8):1015-1025.
[50]周强,甘甫平,王润生,等.高光谱遥感影像矿物自动识别与应用[J],国土资源遥感,2005(4):28-32.
[51]周源,方圣辉,李德仁.利用光谱角敏感森林的高光谱数据快速匹配方法[J].武汉大学学报(信息科学版),2011,(6):687-690.
[52]郑伟,曾志远.遥感图像大气校正方法综述[J].遥感信息,2004(4):66-70.
[53]郑丽.混合像元分解及其应用研究[D].重庆，重庆交通大学，2010.
[54]张兵.时空信息辅助下的高光谱数据挖掘[D].北京:中国科学院遥感应用研究所,2002.
[55]张学工译.统计学习理论的本质[M].北京:清华大学出版社.2000.
[56]张学工.关于统计学习理论与支持向量机[J].自动化学报,2000,26(1):32-42.
[57]张良培,张立福.高光谱遥感[M].武汉:武汉大学出版社,2005.
[58]张莲蓬.基于投影寻踪和非线性主曲线的高光谱遥感影像特征提取及分类研究[D].济南:山东科技大学,2003.
[59]张友静,黄浩,马雪梅.基于KPCA和SAM的城市植被遥感分类研究[J],地理与地理信息科学,2006,22(3):35-38.
[60] Adler-Golden S.M., Matthew M. W., Bernstein L. S., et al. Atmospheric Correction forShortwave Spectral Imagery Based on MODTRAN4[C]. SPIE Proceedings on ImagingSpectrometry,1999,3753:61-69.
[61] Albers J.P.,Stewart J.H. Geology and Mineral Deposits of EsmeraldaCounty,Nevada[C].Nevada Bureau of Mines and Geology Bulletin,1972,78:80.
[62] Ashley R.P., Abrams M.J. Alteration Mapping Using Multispectral Images—Cuprite MiningDistrict, Esmeralda County, Nevada[R].U.S. GeologicalSurvey Open-File Report.1980,17:80-367.
[63] Bach F. R., Jordan M. I. Kernel Independent Component Analysis[J], J. Mach. Learn.Res.,2002,3:1–48.
[64] Bazi Y., Melgani F. Toward an Optimal SVM Classification System for HyperspectralRemote Sensing Images[J]. Geoscience and Remote Sensing,2006,44:3374-3385.
[65] Boardman J.W. Automated Spectral Unmixing Of AVIRIS Data Using Convex GeometryConcepts[C]. Fourth JPL Airborne Geoscience workshop，JPL publication93-26，1993，1:11-14.
[66] Boardman J.W.，Kruse F.A.，and Green R.O. Mapping Target Signatures Via PartialUnmixing of AVIRIS data[C].Fifth JPL Airborne Earth Seience Workshop，JPL Publication95-l，1995，l:23-26.
[67] Boardman J. W. Leveraging the High Dimensionality of AVIRIS Data for ImprovedSub-Pixel Target Unmixing and Rejection of False Positives: Mixture Tuned MatchedFiltering[C]. Summaries of the Seventh JPL Airborne Geoscience Workshop. JPL Publication97-1,1998:55–56.
[68] Boardman J.W. Kruse F.A. Analysis of Imaging Spectrometer Data Using N-DimensionalGeometry and a Mixture Tuned Matched Filtering Approach[J]. IEEE Transactions OnGeoscience And Remote Sensing,2011,1-15.
[69] Camps-Valls G., Bruzzone L.Kernel Methods for Remote Sensing Data Analysis[M]. JohnWiley&Sons, Ltd,2009.
[70] Chen J. Y.,Reed I. S. A Detection Algorithm for Optical Targets in Clutter[J].IEEE Trans.Aerosp. Electron. Syst.,1987, AES-23(1):46–59.
[71] Clark R. N.，Swayze G. A.,Gallagher A. Mapping the Mineralogy and Lithology ofCanyonlands，Utah with Imaging Spectrometer Data and the Multiple Spectral FeatureMapping Algorithm[C]. Third Annual JPL Airborne Geoscience Workshop，JPL Publication92-14，1992,1:11-13.
[72] Clark R. N.，and Swayze G. A. Mapping Minerals，Amorphous Materials，EnvironmentalMaterials，Vegetation，Water，Ice，and Snow，and Other Materials：the USGS TricorderAlgorithm[C]. Fifth Annual JPL Airborne Earth Science Workshop，JPL Publication95-1，1995,39-40.
[73] Desbarats A. J., Dimitrakopoulos R. Geostatistical Simulation of Regionalized Pore-SizeDistributions Using Min/Max Autocorrelation Factors[J], MathematicalGeology,2000,32(8):919-943.
[74] Goetz A. F. Hyperspectral Imaging:Advances in a Spectrum of Applications[C]. Proc the5thAustralian Remote Sensing Conference, Perth,8-12October,1990.
[75] Goetz A. F.,Srivastava V. Mineralogical Mapping in the Cuprite Mining District，Nevada[C].Proceedings of the Airborne Imaging Spectrometer Data Analysis Workshop， JPLPublication85-41，Pasadena，CA，1985,22-29.
[76] Green A. A., Berman M., Switzer P. A Transformation for Ordering Multispectral Data inTerms of Image Quality with Implications for Noise Removal[J]. IEEE Transactions onGeoscience and Remote Sensing,1988,26(1):65-74.
[77] Hsu C.W., Lin C.J. A Comparison of Methods for Multiclass Support Vector Machines[J],IEEE Transactions On Neural Networks,2002,13(2):415-425.
[78] http//www.kernel-methods.net
[79] http://speclab.cr.usgs.gov/
[80] http://aviris.jpl.nasa.gov/html/aviris.freedata.html
[81] Hughes, G.F. On the Mean Accuracy of Statistical Pattern Recognizers [J]. IEEE Transactionon Information Theory,1968,14:55-63.
[82] Kruse F.A.， Raines G.l.,Watson， K. Analytical Techniques for Extracting GeologicInformation from Multichannel Airborne Spectroradiometer and Airborne ImagingSpectrometer Data[C].Proceedings of4th Thematic Conference on Remote Sensing forExploration Geology，Environmental Research Institute of Michigan (ERIM)，Ann Arbor，1985，309-324.
[83] Kruse F. A. Use of Airborne Imaging Spectrometer Data to Map Minerals Associated withHydrothermally Altered Rocks in the Northern Grapevine Mountains， Nevada andCalifornia[J].Remote Sensing of Environment，1988,24(1):31-51.
[84] Kruse F. A.，Kierein-Young K. S., Boardman J. W. Mineral Mapping at Cuprite，Nevadawith a63Channel Imaging Spectrometer[J]. Photogrammetric Engineering and RemoteSensing，1990,56(1):83-92.
[85] Kruse F.A. et al. The Spectral Image Processing System(SIPS)-Interactive Visualization andAnalysis of Imaging Spectrometer Data[J], Remote sensing of Environment，1993，44:145-163.
[86] Lai P. L., Fyfe C. Kernel and Nonlinear Canonical Correlation Analysis[J].Int. J. NeuralSyst.[J],2000,10(5):365-377.
[87] Lawrence C.R.,et al. Mapping Hydrothermally Altered Rocks At Cuprite, Nevada,Using theAdvanced Spaceborne Thermal Emission and Reflection Radiometer (Aster),A New SatelliteImaging System[J]. Economic Geology,2003,98:1019–1027.
[88] Matthew, M.W., Adler-Golden S.M., Berk A., et al. Status of Atmospheric Correction Using aMODTRAN4-based Algorithm[C]. SPIE Proceedings, Algorithms for Multispectral,Hyperspectral, and Ultraspectral Imagery VI.2000,4049:199-207.
[89] Marconcini M., Camps-Valls G., Bruzzone L. A Composite Semisupervised SVM forClassification of Hyperspectral Images[J]. Geoscience and Remote Sensing Letters,2009,6:234-238.
[90] Mazer A.S., Martin M., Lee，M.,et al. Image Proeessing Software for Imaging SpectrometryAnalysis[J]. Remote Sensing of Environment，1988，24(l):201-210.
[91] Mercier G., Lennon M.,et al. Support Vector Machines For Hyperspectral ImageClassification with Spectral-Based Kernels[C]. Geoscience and Remote Sensing Symposium,2003,1:288-290.
[92] Nielsen A.A. Kernel Maximum Autocorrelation Factor and Minimum Noise FractionTransformations[J]. IEEE Transactions On Image Processing,2011,20(3):612-624.
[93] Pal M.,Mather P.M.,Support Vector Machines for Classification in RemoteSensing[J],International Journal of Remote Sensing,2005,26(5):1007-1011.
[94] Plaza A.,Chang C.I.Fast Implementation of Pixel Purity Index Algorithm[C]. Proceedingsof SPIE, Bellingham,2005,5806:307-317.
[95] Plaza A.,Chang C.I. An Improved N-FINDR Algorithm in Implementation[J]. Algorithmsand Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XI,2005,5806:298-306.
[96] Resmini R.G. Mineral mapping with Hyperspeetral Digital Imagery CollectionExperiment(HYDICE) Sensor Data at Cuprite，Nevada，USA[J].Int J Remote Sensing，1997，18:2553-1570.
[97] Roberts D. A.，Yamaguchi Y., Lyon，R. J. P. Comparison of Various Techniques forCalibration of AIS data[C]. Proceedings of2nd AIS workshop，JPL Publication86-35，Pasadena，1986，21-30.
[98] Sch lkopf B., Smola A., Müller K.R. Nonlinear Component Analysis as A KernelEigenvalue Problem[J]. Neural Comput,1998,10(5):1299-1319.
[99] Sch lkopf B.,Smola A. Learning With Kernels[M]. Cambridge, MA:MIT Press,2002.
[100] Switzer P. Green A. A., Min/max Autocorrelation Factors for Multivariate SpatialImagery[R]. Dept. of Statistics, Stanford University.1984.
[101] Winter M. E. N-FINDR: An Algorithm for Fast Autonomous Spectral EndmemberDetermination in Hyperspectral Data[C]. Proceedings of SPIE,1999,3753:266-275.
[102] Winter M.E. Comparison of Approaches for Determining Endmembers in HyperspectralData[C]. Aerospace Conference Proceedings,2000,3:305-313.
[103] Xiu-ping Jia.,Richards John A. Cluster-space Representation for Hyperspectral DataClassification[J]. IEEE Transactions on Geoscience and Remote Sensing,2002,40(3):593-598.
[104] Zhuo L,Zheng J,Wang F,et al. A Genetic Algorithm Based Wrapper Feature SelectionMethod for Classification of Hyperspectral Images Using SVM[C]，The InternationalArchives of the Photogrammetry,Remote Sensing and Spatial InformationSciences.Beijing,2008,XXXVII(B7):397-402.
[2]测绘科技发展“十二五”规划（征求意见稿）.国家测绘地理信息局科技与国际合作司.2010年11月.
[3]杜培军,陈云浩,方涛,等.高光谱遥感数据光谱特征的提取与应用[J].中国矿业大学学报[J],2003,32(5):500-504.
[4]冯静,舒宁.改进型遗传算法和支持向量机的波段选择研究[J],武汉理工大学学报,2009,31(18):120-123.
[5]冯静,舒宁.基于新颖子波变换的高光谱遥感图像特征提取[J].中国图象图形学报,2009,14(10):2042-2046.
[6]甘甫平,王润生,马蔼乃.基于特征谱带的高光谱遥感矿物谱系识别[J].地学前缘(中国地质大学,北京),2003,10(2):445-454.
[7]官莉,刘旸,张雪慧.人工神经网络算法在红外高光谱资料反演大气温度廓线中的应用[J],大气科学学报,2010,33(3):341-346.
[8]耿修瑞，赵永超，刘素红,等.高维义积的矩阵计算以及在高光谱图像端元自动提取中的应用[J].中国科学:信息科学，2010，40:646-652
[9]李建民,张钹,林福宗.支持向量机的训练算法[J].清华大学学报(自然科学版),2003,43(1):120-124.
[10]李素,李文正,周建军,等.遥感影像混合像元分解中的端元选择方法综述[J].地理与地理信息科学,2007,23(5):35-39.
[11]李姗姗,田庆.高光谱遥感图像的端元递进提取算法[J],遥感学报，2009,13(2)：269-275
[12]李庆震,祝小平.基于核PCA的智能图像分析算法[J].弹箭与制导学报,2007,27(5):189-193
[13]刘颖,谷延锋,张晔.基于改进遗传算法的超光谱图像特征选择方法[J].哈尔滨工业大学学报,2005,37(6):733-736.
[14]刘卓,易东云.投影寻踪方法与高光谱遥感图像数据特征提取的研究[J].数学理论与应用,2003,23(1):76-81.
[15]刘正军,王长耀,张继贤.基于小波与遗传算法的特征提取与特征选择[J].遥感学报,2005,9(2):176-185.
[16]刘正军.高维遥感数据土地覆盖特征提取与分类研究[D].北京:中国科学院遥感应用研究所，2003.
[17]刘汉湖.岩矿波谱数据分析与信息提取方法研究[D].成都:成都理工大学,2008.
[18]刘翔.基于光谱维变换的高光谱图像目标探测研究[D].北京:中国科学院研究生院,2008.
[19]刘江永,王大明.基于支持向量机的快速高光谱分类研究[J],陕西师范大学学报(自然科学版),2009,37(4):43-47.
[20]骆剑承,周成虎,梁怡,等.支撑向量机及其遥感影像空间特征提取和分类的应用研究[J].遥感学报,2002,6(1):50-55.
[21]梁顺林.定量遥感[M],北京:科学出版社,2009.
[22]路威.面向目标探测的高光谱影像特征提取与分类技术研究[D].郑州:解放军信息工程大学,2005.
[23]林娜,杨武年,刘汉湖.基于高光谱遥感的岩矿端元识别及信息提取研究[J].遥感信息,2011,5:114-118.
[24]牟少敏.核方法的研究及其应用[D].北京：北方交通大学，2008.
[25]梅安新，彭望琭，秦其明,等.遥感导论[M].北京:高等教育出版社,2001.
[26]牛鹏,魏维.基SVM动态集成的高光谱遥感图像分类[J].计算机应用,2010,30(6):1590-1593.
[27]佘红伟，张艳宁，袁和金.一种无监督高光谱图像分类算法[J].中国图像图形学报,2008,13(6):1123-1128.
[28]苏红军,杜培军.高光谱数据特征选择与特征提取研究[J].遥感技术与应用，2006,21(4):288-293.
[29]沈照庆,陶建斌.基于模糊核主成分分析的高光谱遥感影像特征提取研究[J].国土资源遥感,2009,(3):41-45.
[30]谭琨,杜培军.基于再生核Hilbert空间小波核函数支持向量机的高光谱遥感影像分类[J].测绘学报,2011,40(2):142-147.
[31]童庆禧,张兵,郑兰芬.高光谱遥感:原理、技术与应用[M].北京:高等教育出版社,2006.
[32]田庆久,宫鹏.地物波谱数据库研究现状与发展趋势[J].遥感信息，2002，3：1-6.
[33]王润生,甘甫平,闫柏琨.高光谱矿物填图技术与应用研究[J].国土资源遥感,2010,83(1):1-13.
[34]万家强,王越,刘羽.改进KPCA对分类数据的特征提取[J],计算机工程与设计，2010，31（18）：4085-4088.
[35]王国胜.核函数的性质及其构造方法[J].计算机科学，2006，33(6):172-178.
[36]王毅,张良培,李平湘.基于自动搜索和光谱匹配技术的训练样本纯化算法[J].武汉大学学报(信息科学版),2007(3):216-219.
[37]吴波,张良培,李平湘.基于支撑向量回归的高光谱混合像元非线性分解[J].遥感学报,2006,10(3):312-318.
[38]吴波,张良培,李平湘.高光谱端元自动提取的迭代分解方法[J],遥感学报,2005,(3).
[39]熊桢,童庆禧,郑兰芬.用于高光谱遥感图像分类的一种高阶神经网络算法[J].中国图象图形学报,2000,(3)196-201.
[40]徐宏根,马洪超,李德仁.结合SOM神经网络和混合像元分解的高光谱影像分类方法研究[J].遥感学报,2007,11(6):778-786.
[41]徐芳.航空影像分割的支持向量机方法[D].武汉:武汉大学,2004.
[42]徐元进.面向找矿的高光谱遥感岩矿信息提取方法研究[D].武汉:中国地质大学,2009.
[43]薛宁静.多类支持向量机分类器对比研究[J].计算机工程与设计,2011,32(5):1792-1795
[44]闰永忠,万余庆.高光谱图像模糊识别分类及其精度评价[J].地球信息科学,2005,7(4):20-24.
[45]杨诸胜,郭雷,罗欣,等.一种基于主成分分析的高光谱图像波段选择算法[J].微电子学与计算机,2006,23(12):72-74.
[46]杨飞,张柏,刘志明,等.玉米冠层FPAR的高光谱遥感估算研究——基于PCA方法及近、短波红外波段[J].国土资源遥感,2008,(4):9-13.
[47]杨国鹏.基于核方法的高光谱影像分类与特征提取[D],郑州：解放军信息工程大学，2007.
[48]杨国鹏.基于机器学习方法的高光谱影像分类研究[D].河南:解放军信息工程大学,2010.
[49]尹峻松,肖健,周宗潭,等.非线性流形学习方法的分析与应用[J],自然科学进展,2007,17(8):1015-1025.
[50]周强,甘甫平,王润生,等.高光谱遥感影像矿物自动识别与应用[J],国土资源遥感,2005(4):28-32.
[51]周源,方圣辉,李德仁.利用光谱角敏感森林的高光谱数据快速匹配方法[J].武汉大学学报(信息科学版),2011,(6):687-690.
[52]郑伟,曾志远.遥感图像大气校正方法综述[J].遥感信息,2004(4):66-70.
[53]郑丽.混合像元分解及其应用研究[D].重庆，重庆交通大学，2010.
[54]张兵.时空信息辅助下的高光谱数据挖掘[D].北京:中国科学院遥感应用研究所,2002.
[55]张学工译.统计学习理论的本质[M].北京:清华大学出版社.2000.
[56]张学工.关于统计学习理论与支持向量机[J].自动化学报,2000,26(1):32-42.
[57]张良培,张立福.高光谱遥感[M].武汉:武汉大学出版社,2005.
[58]张莲蓬.基于投影寻踪和非线性主曲线的高光谱遥感影像特征提取及分类研究[D].济南:山东科技大学,2003.
[59]张友静,黄浩,马雪梅.基于KPCA和SAM的城市植被遥感分类研究[J],地理与地理信息科学,2006,22(3):35-38.
[60] Adler-Golden S.M., Matthew M. W., Bernstein L. S., et al. Atmospheric Correction forShortwave Spectral Imagery Based on MODTRAN4[C]. SPIE Proceedings on ImagingSpectrometry,1999,3753:61-69.
[61] Albers J.P.,Stewart J.H. Geology and Mineral Deposits of EsmeraldaCounty,Nevada[C].Nevada Bureau of Mines and Geology Bulletin,1972,78:80.
[62] Ashley R.P., Abrams M.J. Alteration Mapping Using Multispectral Images—Cuprite MiningDistrict, Esmeralda County, Nevada[R].U.S. GeologicalSurvey Open-File Report.1980,17:80-367.
[63] Bach F. R., Jordan M. I. Kernel Independent Component Analysis[J], J. Mach. Learn.Res.,2002,3:1–48.
[64] Bazi Y., Melgani F. Toward an Optimal SVM Classification System for HyperspectralRemote Sensing Images[J]. Geoscience and Remote Sensing,2006,44:3374-3385.
[65] Boardman J.W. Automated Spectral Unmixing Of AVIRIS Data Using Convex GeometryConcepts[C]. Fourth JPL Airborne Geoscience workshop，JPL publication93-26，1993，1:11-14.
[66] Boardman J.W.，Kruse F.A.，and Green R.O. Mapping Target Signatures Via PartialUnmixing of AVIRIS data[C].Fifth JPL Airborne Earth Seience Workshop，JPL Publication95-l，1995，l:23-26.
[67] Boardman J. W. Leveraging the High Dimensionality of AVIRIS Data for ImprovedSub-Pixel Target Unmixing and Rejection of False Positives: Mixture Tuned MatchedFiltering[C]. Summaries of the Seventh JPL Airborne Geoscience Workshop. JPL Publication97-1,1998:55–56.
[68] Boardman J.W. Kruse F.A. Analysis of Imaging Spectrometer Data Using N-DimensionalGeometry and a Mixture Tuned Matched Filtering Approach[J]. IEEE Transactions OnGeoscience And Remote Sensing,2011,1-15.
[69] Camps-Valls G., Bruzzone L.Kernel Methods for Remote Sensing Data Analysis[M]. JohnWiley&Sons, Ltd,2009.
[70] Chen J. Y.,Reed I. S. A Detection Algorithm for Optical Targets in Clutter[J].IEEE Trans.Aerosp. Electron. Syst.,1987, AES-23(1):46–59.
[71] Clark R. N.，Swayze G. A.,Gallagher A. Mapping the Mineralogy and Lithology ofCanyonlands，Utah with Imaging Spectrometer Data and the Multiple Spectral FeatureMapping Algorithm[C]. Third Annual JPL Airborne Geoscience Workshop，JPL Publication92-14，1992,1:11-13.
[72] Clark R. N.，and Swayze G. A. Mapping Minerals，Amorphous Materials，EnvironmentalMaterials，Vegetation，Water，Ice，and Snow，and Other Materials：the USGS TricorderAlgorithm[C]. Fifth Annual JPL Airborne Earth Science Workshop，JPL Publication95-1，1995,39-40.
[73] Desbarats A. J., Dimitrakopoulos R. Geostatistical Simulation of Regionalized Pore-SizeDistributions Using Min/Max Autocorrelation Factors[J], MathematicalGeology,2000,32(8):919-943.
[74] Goetz A. F. Hyperspectral Imaging:Advances in a Spectrum of Applications[C]. Proc the5thAustralian Remote Sensing Conference, Perth,8-12October,1990.
[75] Goetz A. F.,Srivastava V. Mineralogical Mapping in the Cuprite Mining District，Nevada[C].Proceedings of the Airborne Imaging Spectrometer Data Analysis Workshop， JPLPublication85-41，Pasadena，CA，1985,22-29.
[76] Green A. A., Berman M., Switzer P. A Transformation for Ordering Multispectral Data inTerms of Image Quality with Implications for Noise Removal[J]. IEEE Transactions onGeoscience and Remote Sensing,1988,26(1):65-74.
[77] Hsu C.W., Lin C.J. A Comparison of Methods for Multiclass Support Vector Machines[J],IEEE Transactions On Neural Networks,2002,13(2):415-425.
[78] http//www.kernel-methods.net
[79] http://speclab.cr.usgs.gov/
[80] http://aviris.jpl.nasa.gov/html/aviris.freedata.html
[81] Hughes, G.F. On the Mean Accuracy of Statistical Pattern Recognizers [J]. IEEE Transactionon Information Theory,1968,14:55-63.
[82] Kruse F.A.， Raines G.l.,Watson， K. Analytical Techniques for Extracting GeologicInformation from Multichannel Airborne Spectroradiometer and Airborne ImagingSpectrometer Data[C].Proceedings of4th Thematic Conference on Remote Sensing forExploration Geology，Environmental Research Institute of Michigan (ERIM)，Ann Arbor，1985，309-324.
[83] Kruse F. A. Use of Airborne Imaging Spectrometer Data to Map Minerals Associated withHydrothermally Altered Rocks in the Northern Grapevine Mountains， Nevada andCalifornia[J].Remote Sensing of Environment，1988,24(1):31-51.
[84] Kruse F. A.，Kierein-Young K. S., Boardman J. W. Mineral Mapping at Cuprite，Nevadawith a63Channel Imaging Spectrometer[J]. Photogrammetric Engineering and RemoteSensing，1990,56(1):83-92.
[85] Kruse F.A. et al. The Spectral Image Processing System(SIPS)-Interactive Visualization andAnalysis of Imaging Spectrometer Data[J], Remote sensing of Environment，1993，44:145-163.
[86] Lai P. L., Fyfe C. Kernel and Nonlinear Canonical Correlation Analysis[J].Int. J. NeuralSyst.[J],2000,10(5):365-377.
[87] Lawrence C.R.,et al. Mapping Hydrothermally Altered Rocks At Cuprite, Nevada,Using theAdvanced Spaceborne Thermal Emission and Reflection Radiometer (Aster),A New SatelliteImaging System[J]. Economic Geology,2003,98:1019–1027.
[88] Matthew, M.W., Adler-Golden S.M., Berk A., et al. Status of Atmospheric Correction Using aMODTRAN4-based Algorithm[C]. SPIE Proceedings, Algorithms for Multispectral,Hyperspectral, and Ultraspectral Imagery VI.2000,4049:199-207.
[89] Marconcini M., Camps-Valls G., Bruzzone L. A Composite Semisupervised SVM forClassification of Hyperspectral Images[J]. Geoscience and Remote Sensing Letters,2009,6:234-238.
[90] Mazer A.S., Martin M., Lee，M.,et al. Image Proeessing Software for Imaging SpectrometryAnalysis[J]. Remote Sensing of Environment，1988，24(l):201-210.
[91] Mercier G., Lennon M.,et al. Support Vector Machines For Hyperspectral ImageClassification with Spectral-Based Kernels[C]. Geoscience and Remote Sensing Symposium,2003,1:288-290.
[92] Nielsen A.A. Kernel Maximum Autocorrelation Factor and Minimum Noise FractionTransformations[J]. IEEE Transactions On Image Processing,2011,20(3):612-624.
[93] Pal M.,Mather P.M.,Support Vector Machines for Classification in RemoteSensing[J],International Journal of Remote Sensing,2005,26(5):1007-1011.
[94] Plaza A.,Chang C.I.Fast Implementation of Pixel Purity Index Algorithm[C]. Proceedingsof SPIE, Bellingham,2005,5806:307-317.
[95] Plaza A.,Chang C.I. An Improved N-FINDR Algorithm in Implementation[J]. Algorithmsand Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XI,2005,5806:298-306.
[96] Resmini R.G. Mineral mapping with Hyperspeetral Digital Imagery CollectionExperiment(HYDICE) Sensor Data at Cuprite，Nevada，USA[J].Int J Remote Sensing，1997，18:2553-1570.
[97] Roberts D. A.，Yamaguchi Y., Lyon，R. J. P. Comparison of Various Techniques forCalibration of AIS data[C]. Proceedings of2nd AIS workshop，JPL Publication86-35，Pasadena，1986，21-30.
[98] Sch lkopf B., Smola A., Müller K.R. Nonlinear Component Analysis as A KernelEigenvalue Problem[J]. Neural Comput,1998,10(5):1299-1319.
[99] Sch lkopf B.,Smola A. Learning With Kernels[M]. Cambridge, MA:MIT Press,2002.
[100] Switzer P. Green A. A., Min/max Autocorrelation Factors for Multivariate SpatialImagery[R]. Dept. of Statistics, Stanford University.1984.
[101] Winter M. E. N-FINDR: An Algorithm for Fast Autonomous Spectral EndmemberDetermination in Hyperspectral Data[C]. Proceedings of SPIE,1999,3753:266-275.
[102] Winter M.E. Comparison of Approaches for Determining Endmembers in HyperspectralData[C]. Aerospace Conference Proceedings,2000,3:305-313.
[103] Xiu-ping Jia.,Richards John A. Cluster-space Representation for Hyperspectral DataClassification[J]. IEEE Transactions on Geoscience and Remote Sensing,2002,40(3):593-598.
[104] Zhuo L,Zheng J,Wang F,et al. A Genetic Algorithm Based Wrapper Feature SelectionMethod for Classification of Hyperspectral Images Using SVM[C]，The InternationalArchives of the Photogrammetry,Remote Sensing and Spatial InformationSciences.Beijing,2008,XXXVII(B7):397-402.