基于数据融合的高光谱遥感图像分类研究
|
摘要
高光谱遥感技术的快速发展,使得获取更高光谱分辨率的地面信息成为可能,为定量遥感的实现创造了有利条件。然而,高光谱数据具有波段数目众多、各波段相关性强、运算量大的特点,这对相应的处理技术提出了很高的要求。高光谱遥感影像分类是高光谱遥感技术的一个重要应用,针对多光谱遥感图像的分类人们已经研究了多种处理方法,技术比较成熟;但是高光谱图像的数据量大,数据维高,使得通常的多光谱图像处理方法对于高光谱图像的应用有较大的限制。为了解决这一问题,本文在深入分析高光谱图像数据特点的基础上,重点研究了基于数据融合的高光谱遥感影像分类技术,研究的主要内容如下:
首先,将二进脊波变换应用于高光谱遥感图像的数据融合,并结合二进脊波变换数据分解的特点,提出了相应的融合策略,实现了数据级融合。该算法先对同一波段组内的各个子图像进行有限Randon变换,将线性奇异转化为点奇异;然后通过二进小波变换对点奇异信号进行处理。在融合策略的选取中,充分考虑到小波变换进行数据分解的特点:对于代表图像概貌信息的低频部分采用归一化方差加权融合;对于包含图像细节和纹理特征的高频部分选择像素绝对值最大的部分作为融合后的像素值。在尽可能多地保持原始图像信息的前提下,实现了对AVIRIS图像的像素级融合,并在此基础上进行了地物分类。仿真实验表明,该方案能有效地改善融合效果,并进一步提升分类精度。
其次,针对有限脊波变换存在的“环绕效应”所引入的噪声,研究了减小其影响的方案。研究表明:图像分割子块尺寸越大,“环绕效应”的影响就越大;子块尺寸越小,脊波的优势就越明显。但是,分割尺寸越小,重构图像的块状效应越明显,而且分割越小图像所表示的方向也越少,效果也近似于小波变换。因此,在实际选择中应该折衷考虑。
再次,为了彻底消除“环绕效应”,研究了基于真实脊函数和快速SlantStack算法的数字脊波变换在高光谱图像融合中的应用。由于没有采用有限Randon变换实现脊波的数字化,它能彻底消除“环绕效应”,融合的效果也能得到进一步改善,但是引入了数据冗余。为进一步提高分类精度与运算速度,本文将其与一种新的神经网络——样条权函数神经网络相结合,实现了特征级融合分类。由于该神经网络所具有的对初值不敏感、收敛速度快和不存在局部极小等优点,该方案能得到较高的分类精度。
最后,研究了数字脊波和样条权函数神经网络在高光谱图像决策级融合中的应用,对二次融合进行了尝试。该方案先采用数字脊波实现像素级融合,并利用样条权神经网络实现局部分类;然后结合主体投票规则将各局部分类器输出的结果进行决策融合。仿真实验结果表明,该方案能在较少的训练样本条件下获得较高的分类精度,该决策融合方案要优于基于BP、RBF神经网络实现的决策融合方案。
The rapid development of the hyperspectral remote sensing technology, which made it possible to acquire object information on the surface of the Earth, is very helpful for achievement of the quantitative analysis on remote sensing. Hyperspectral remote sensing imagery generally consists of dozens or hundreds of narrow, also contiguous, spectral bands, which accounts for the computational burden and the phenomenon where the response of bands tends to be highly correlated. Consequently, advanced techniques are needed to exploit the extensive information contained in hyperspectral data. Classification of hyperspectral remote sensing imagery plays an important role in its application. These years, a large amount of algorithm on classifying multispectral data is accomplished by researchers, but the characteristics that hyperspectral data possessed restricts its application on hyperspetral imagery owing to a huge computational burden. To solve this problem, this dissertation focuses on hyperspectral remote sensing imagery classification methodology based on information fusion algorithm after a thorough study on the characteristics of hyperspectral data. The major research are as follows:
Firstly, dyadic ridgelet transform is brought in to achieve information fusion on hyperspectral remote sensing imagery, and a brand new fusion algorithm is put forward which based on the feature of the data gained after ridgelet transform. This method applies finite Randon transform to all the sub images which were classified into the same band set at first, which can change line singularity in the image into point singularity, then dyadic wavelet transform is implemented to deal with the point singular data. While choosing fusion algorithm, the data character gained by wavelet transform is taken into consideration: normal variance is weighed on those data which represents the outline information of the image; as to those data that contains details and texture information, the pixels with biggest absolute value is chosen to represent the pixel value of fused image. After pixel level fusion to AVIRIS image was accomplished, texture classification is done based on the fused image. Experimental results show that this method can effectively improve fusion result, and achieve high overall accuracy of classification.
Secondly, further study is processed to allay the influence of "wraparound effect" brought in by finite ridgelet transform. The study shows that while the larger size of the sub image gained by segmentation of image, the "wraparound effect" takes more important role; while the smaller, the more advantage that ridgelet can embodiment. Nevertheless, the smaller size of sub image also results to clearer block effect, and the fusion results achieved by ridgelet will also approximate to wavelet transform. Thus, compromise should be made due to different consideration while choosing the size of segmentation block.
Thirdly, a kind of digital ridgelet transform based on true ridge functions and fast slant stack algorithm is researched on its application in hyperspectral data fusion in order to eliminate "wraparound effect". Since this algorithm avoid using finite randon transform to discretize ridgelet, this digital ridgelet transform can thoroughly eliminate "wraparound effect", thus achieving better fusion results, but it also brings in data redundancy. In order to improve classification accuracy and computational speed, a new neural network called sample weight neural network (SWNN) was combined with this digital ridgelet. Because of the advantages which SWNN possessed (such as, not so sensitive to initial value, no partial minimum value, with a rapid convergence speed, et al), the combination strategy can achieve higher accuracy and less computational burden than other neural network such as BP neural network, RBF neural network.
Finally, fusion of hyperspectral data on decision level is studied based on digital ridgelet, sample weight neural network and majority voting rule algorithm. In this algorithm, pixel level fusion is achieved by using digital ridegelet transform, then local classification is processed by using SWNN; then decision fusion is attained by synthesize the results of all the local classifiers with majority voting rules. Experimental results show that this algorithm can achieve high classification accuracy even with very limited training samples, and the decision level fusion algorithm based on SWNN network is better than traditional neural networks such as BP, RPF.
首先,将二进脊波变换应用于高光谱遥感图像的数据融合,并结合二进脊波变换数据分解的特点,提出了相应的融合策略,实现了数据级融合。该算法先对同一波段组内的各个子图像进行有限Randon变换,将线性奇异转化为点奇异;然后通过二进小波变换对点奇异信号进行处理。在融合策略的选取中,充分考虑到小波变换进行数据分解的特点:对于代表图像概貌信息的低频部分采用归一化方差加权融合;对于包含图像细节和纹理特征的高频部分选择像素绝对值最大的部分作为融合后的像素值。在尽可能多地保持原始图像信息的前提下,实现了对AVIRIS图像的像素级融合,并在此基础上进行了地物分类。仿真实验表明,该方案能有效地改善融合效果,并进一步提升分类精度。
其次,针对有限脊波变换存在的“环绕效应”所引入的噪声,研究了减小其影响的方案。研究表明:图像分割子块尺寸越大,“环绕效应”的影响就越大;子块尺寸越小,脊波的优势就越明显。但是,分割尺寸越小,重构图像的块状效应越明显,而且分割越小图像所表示的方向也越少,效果也近似于小波变换。因此,在实际选择中应该折衷考虑。
再次,为了彻底消除“环绕效应”,研究了基于真实脊函数和快速SlantStack算法的数字脊波变换在高光谱图像融合中的应用。由于没有采用有限Randon变换实现脊波的数字化,它能彻底消除“环绕效应”,融合的效果也能得到进一步改善,但是引入了数据冗余。为进一步提高分类精度与运算速度,本文将其与一种新的神经网络——样条权函数神经网络相结合,实现了特征级融合分类。由于该神经网络所具有的对初值不敏感、收敛速度快和不存在局部极小等优点,该方案能得到较高的分类精度。
最后,研究了数字脊波和样条权函数神经网络在高光谱图像决策级融合中的应用,对二次融合进行了尝试。该方案先采用数字脊波实现像素级融合,并利用样条权神经网络实现局部分类;然后结合主体投票规则将各局部分类器输出的结果进行决策融合。仿真实验结果表明,该方案能在较少的训练样本条件下获得较高的分类精度,该决策融合方案要优于基于BP、RBF神经网络实现的决策融合方案。
The rapid development of the hyperspectral remote sensing technology, which made it possible to acquire object information on the surface of the Earth, is very helpful for achievement of the quantitative analysis on remote sensing. Hyperspectral remote sensing imagery generally consists of dozens or hundreds of narrow, also contiguous, spectral bands, which accounts for the computational burden and the phenomenon where the response of bands tends to be highly correlated. Consequently, advanced techniques are needed to exploit the extensive information contained in hyperspectral data. Classification of hyperspectral remote sensing imagery plays an important role in its application. These years, a large amount of algorithm on classifying multispectral data is accomplished by researchers, but the characteristics that hyperspectral data possessed restricts its application on hyperspetral imagery owing to a huge computational burden. To solve this problem, this dissertation focuses on hyperspectral remote sensing imagery classification methodology based on information fusion algorithm after a thorough study on the characteristics of hyperspectral data. The major research are as follows:
Firstly, dyadic ridgelet transform is brought in to achieve information fusion on hyperspectral remote sensing imagery, and a brand new fusion algorithm is put forward which based on the feature of the data gained after ridgelet transform. This method applies finite Randon transform to all the sub images which were classified into the same band set at first, which can change line singularity in the image into point singularity, then dyadic wavelet transform is implemented to deal with the point singular data. While choosing fusion algorithm, the data character gained by wavelet transform is taken into consideration: normal variance is weighed on those data which represents the outline information of the image; as to those data that contains details and texture information, the pixels with biggest absolute value is chosen to represent the pixel value of fused image. After pixel level fusion to AVIRIS image was accomplished, texture classification is done based on the fused image. Experimental results show that this method can effectively improve fusion result, and achieve high overall accuracy of classification.
Secondly, further study is processed to allay the influence of "wraparound effect" brought in by finite ridgelet transform. The study shows that while the larger size of the sub image gained by segmentation of image, the "wraparound effect" takes more important role; while the smaller, the more advantage that ridgelet can embodiment. Nevertheless, the smaller size of sub image also results to clearer block effect, and the fusion results achieved by ridgelet will also approximate to wavelet transform. Thus, compromise should be made due to different consideration while choosing the size of segmentation block.
Thirdly, a kind of digital ridgelet transform based on true ridge functions and fast slant stack algorithm is researched on its application in hyperspectral data fusion in order to eliminate "wraparound effect". Since this algorithm avoid using finite randon transform to discretize ridgelet, this digital ridgelet transform can thoroughly eliminate "wraparound effect", thus achieving better fusion results, but it also brings in data redundancy. In order to improve classification accuracy and computational speed, a new neural network called sample weight neural network (SWNN) was combined with this digital ridgelet. Because of the advantages which SWNN possessed (such as, not so sensitive to initial value, no partial minimum value, with a rapid convergence speed, et al), the combination strategy can achieve higher accuracy and less computational burden than other neural network such as BP neural network, RBF neural network.
Finally, fusion of hyperspectral data on decision level is studied based on digital ridgelet, sample weight neural network and majority voting rule algorithm. In this algorithm, pixel level fusion is achieved by using digital ridegelet transform, then local classification is processed by using SWNN; then decision fusion is attained by synthesize the results of all the local classifiers with majority voting rules. Experimental results show that this algorithm can achieve high classification accuracy even with very limited training samples, and the decision level fusion algorithm based on SWNN network is better than traditional neural networks such as BP, RPF.
引文
[1]张良培,张立福.高光谱遥感.武汉:武汉大学出版社,2005:1-10页,58-126页
[2]童庆禧,张兵,郑兰芬.高光谱遥感——原理、技术与应用.北京:高等教育出版社,2006:166-218页,293-316页
[3]浦瑞良,宫鹏.高光谱遥感及其应用.北京:高等教育出版社,2000:1-45页,123-228页
[4]徐青,张艳,耿则勋,邢帅,谭兵.遥感影像融合与分辨率增强技术.北京:科学出版社,2007:68-101页
[5]边肇祺,张学工.模式识别.北京:清华大学出版社,1999
[6]刘春红.超光谱遥感图像降维及分类方法研究.哈尔工程大学博士论文,2005:1-20页
[7]马建文,李启青,哈斯巴干,戴芹.遥感数据智能处理方法与程序设计.北京:科学出版社,2005
[8]Lee Sanghoon,Crawford Melba M.Hierarchical Clustering Approach for Unsupervised Image Classification of Hyperspectral Data.International Geoscience and Remote Sensing Symposium(IGARSS).2004,2:941-944 P
[9]Jimenez Luis O,Rivera-Medina Jorge L.Integration of spatial and Spectral Information by means of Unsupervised Extraction and Classification for Homogenous Objects Applied to Multispectral Data.IEEE Transactions on Geosciences and Remote Sensing.2005,43(4):844-851 P
[10]Du Qian.Independent Component Analysis to Hyperspectral Image Classification.Proceedings of SPIE.-The International Society for Optical Engineering.Imaging Spectrometry X.2004,5546:366-373 P
[11]Shah Chintan A,Arora Manoj K,Varshey Pramod K.Unsupervised Classification of Hyperspectral data:An ICA Mixture Model based Approach. International Journal of Remote Sensing. 2004,25 (2) :481-487 P
[12] Dudar M.M, Landgrebe D.A. Toward an Optimal supervised Classifier for the Analysis of Hyperspectral data. IEEE Transactions on Geoscience and Remote Sensing. 2004, 42 (1) : 271-277 P
[13] Rand Robert S. Spectral/Spatial Annealing of Hyperspctral Imagery Initialized by a Supervised Classification Method. Proceeding of SPIE—the International Society for Optical Engineering. 2002, 4725:242-253 P
[14] Landgrebe David, Murat Dunder M. A Model-based Mixture-supervised Classification Approach in Hyperspectral Data Analysis. IEEE Transactions on Geoscience and Remote Sensing. 2002, 40 (12) :2692-2699 P
[15] Chang Yang-Lang, Han Chin-Chuan. Greedy Modular Eigenspaces and Positive Boolean Function for Supervised Hyperspectral Image Classification. Optical Engineering. 2003, 42 (9): 2576-2587 P
[16] Rosario Dalton. A Semiparamrtric Approach Using Discriminant Metric SAM (Spectral Angle Mapper). Proceeding of SPIE—The International Society for Optical Engineering. 2004, 5426:58-66 P
[17] Goetz Alexandar F.H, Vane Gregg, et al. Image Spectrometry for Earth Remote Sensing. Science. 1985, 228 (4704) : 1147-1153 P
[18] Schmidt K.S, Skidmore A.K, et al. Mapping Coastal Vegetation Using an Expert System and Hyperspectral Imagery. Photogram metric Engineering and Remote Sensing. 2004, 70 (6) : 703-715 P
[19] Gustavo Camps-Vails, Lorenzo Bruzzone. Kernel-Based Methods for Hyperspectral Image Classification. IEEE Transactions on Geoscience and Remote Sensing. Vol.43, No.6, 2005, 7:1351-1362 P
[20] Jimenez L O, et al. Classification of Hyperdimensional Data Based on Feature and Decision Fusion Approaches Using Projection Pursuit,Majority Voting, and Neural Networks. IEEE on G.R.S. 1999, 37(3):1360-1366 P
[21] Jon Atli Benediktsson, Ioannis Kanellopoulos. Classification of Multisource and Hyperspectral Data Based on Decision Fusion. IEEE Trans on Geoscience and Remote Sensing. 1999, 37 (3) : 1367-1377 P
[22] Baofeng Guo, Steve Gunn, Bob Damper, James Nelson. Hyperspectral Image Fusion using Spectrally Weighted Kernels. 7th International conference on Information Fusion. 2005:402-408 P
[23] Anil Cheriyadat, Lori Mann Bruce, et al. Decision level fusion with best-Bases for Hyperspectral Classification. IEEE Trans on Geo science and Remote Sensing. 2004:399-406 P
[24] Jorge Nunez, et al. Multiresolusion-Based Image Fusion with Additive Decomposition. IEEE Trans on Geo science and Remote Sensing, 1999,37 (3): 1204-1211 P
[25] Chunhong Liu, Chunhui Zhao, WAnhai Chen. Hyperspectral Image Classification by Second Generation Wavelet Based on Adaptive Band Selection. Proceedings of the IEEE International Conference on Mechatronics & Automation. 2005:1175-1179 P
[26] Terry A Wilson, Steven K Rogers, Matthew Kabrisky. Perpetual-Based Image Fusion for Hyperspectral Data. IEEE Trans on Geo science and Remote Sensing, 1997, 35 (4) : 1007-1017 P
[27] Jon Atli Benediktsson, Jon Aevar Palmason, et al. Decision Level Fusion in Classification of Hyperspectral Data from Urban Areas.IEEE Trans on Geo science and Remote Sensing. 2004:73-76 P
[28] P.C. Smits. Comparsion of some Feature Subset Selection Methods for Use in Remote Sensing Image Analysis. IEEE Trans on Geo science and Remote Sensing. 2001:530-533 P
[29] Bor-Chen Kuo, et al. Hyperspectral Data Calssification Using Classifier Overproduction and Fusion Strategies. IEEE Trans, 2004:2937-2940 P
[30] Y. Allard, A. Jouan. Interpretation of high-resolution optical imagery with evidential fusion of spectral information and object attributes.9th International Conference on Information Fusion,2006:1-7 P
[31]Boriana L.Milenova,Marcos M.Campos.Mining High-Dimensional Data for Information Fusion:A Database-centric Approach.7th International Conference on Information Fusion.2005:638-645 P
[32]Junping Zhang,Ye Zhang.Fusion Classification of Hyper-spectral Image Based on Adaptive Subspace Decomposition.International Conference on Image Processing,Vol.3,2000:10-13 P
[33]张钧萍,张晔.基于多特征多分辨率融合的高管够图像分类.红外与毫米波学报,2004,Vol.23.No.5:345-348页
[34]Qiang Wang,Yi Shen.Performances Evaluation of Image Fusion Techniques Based on Nonlinear Correlation Measurement.Instrumentation and Measurement Technology Conference.2004:472-475 P
[35]Liang Sun,Chongzhao Han,Ming Lei.Knowledge discovery-based multiple classifier fusion:a Generalized Rough Set Method.9th International Conference on Information Fusion.2006:1-8 P
[36]孙亮,韩崇昭,康欣.关于集值特征属性的知识发现方法研究.Proceedings of the 5th World Congress On Intelligent Control and Automation.2004:3051-3054 P
[37]Joseph C Harsanyi,Chein-IChang.Hyperspectral Image Classification and Dimensionality Reduction:An Orthogonal Subspace Projection Approach.IEEE Trans on G.R.S,1994,32(4):779-785 P
[38]孙延奎.小波分析与应用.北京:机械工业出版社,2005
[39]赵英时.遥感应用分析原理与方法.北京:科学出版社,2003:190-208页
[40]David Landgrebe.Multispectral Data Analysis:A Signal Theory Perspective.http://cobweb.ecn.purdue.edu/%TEbiehl/MultiSpec/
[41]阎平凡,张长水.人工神经网络与模拟进化计算.北京:清华大学出版社,2005
[42]孙即祥.现代模式识别.长沙:国防科技大学出版社,2002
[43]张代远.神经网络新理论与方法.北京:清华大学出版社,2006:51-113页
[44]Landgrebe D.Information Extraction Principles and Methods for Multispectral and Hyperspectral Image Data,Chapter 1 of Information Processing for Remote Sensing,edited by C.H.Chen,published by the World Scientific Publishing Co.Inc.,1060 Main Street,River Edge,NJ07661,USA,2000
[45]Jimenez Luis O,Rivera-Medina Jorge L.Integration of spatial and spectral information by means of unsupervised extraction and classification for homogenous objects applied to multispectral and hyperspectral data.IEEE Transactions on Geoscience and Remote Sensing.2005,43(4):844-851 P
[46]Hughes G..F.On the mean accuracy of statistical pattern recognizers.IEEE Trans.On Information Theory,Jan 1968,Vol.IT-14,No.1
[47]李弼程,魏俊,彭天强.遥感影象融合效果的客观分析与评价.计算机工程与科学,2004,Vol.26,No.1:42-46页
[48]崔岩梅,倪国强.利用统计特性进行图像融合效果分析及评价.北京理工大学学报,2000,20(1):102-146页
[49]E.J.Candes.Ridgelets:Theory and Application.The Department of Statistics and the Committee on Graduate Studies of Stanford University.Stanford University.1998
[50]黄建平,王卫卫.基于二进脊波变换的图像去噪.计算机工程与应用,2006,05:69-79页
[51]Starck J.L,Candes E.J.,Donoho D.L.The curvelet transform for image denoising.IEEE Trans.IP.2002,11(6):670-684 P
[52]吴昊.高光谱遥感图像数据分类技术研究.国防科学技术大学博士学位论文,2004
[53]Candes E.J.,Donoho D.L.Curvelets,Multiresolution Representation,and Scaling Laws.SPIE Wavelet Application in Signal and Image Processing ⅤⅢ,2002,Vol.4119
[54]刘春红,赵春晖.基于提升算法的超谱遥感图像融合分类研究.哈尔滨工程大学学报,2004,Vol.25(No.6):794-798页
[55]Tao Chen,Junping Zhang.Remote Sensing Image Fusion Based on Ridgelet Transform.Geoscience and Remote Sensing Symposium,2005:1150-1153 P
[56]A.Averbuch,R.R.Coifman,D.L.Donoho.Fast Slant Stack:A notion of Radon Transform for Data in a Cartesian Grid which is Rapidly Computible,Algebraically Exact,Geometrically Faithful and Invertible.http://stat.stanford.edu/～donoho/reports
[57]Donoho D.L,Flesia A.G.Digital Ridgelet Transform based on True Ridge Functions.http://stat.stanford.edu/～donoho/reports
[58]白键,冯象初.基于局部对偶框架的数字脊波重构.中国科学(E辑)信息科学,2005,35(10):1072-1082页
[59]张代远.样条权函数神经网络的一种新算法.系统工程与电子技术,2006(9):1434-1437页
[60]张钧萍,张晔,周显廷.基于信息融合的超谱图像分类方法研究.哈尔滨工业大学学报,2002,34(4):464-468页
[61]Zhang Ye.Adaptive subspace decomposition for hyperspectral data dimensionality reduction.ICIP99 Proceedings.1999:326-329 P
[62]Minh N.Do,Martin Vetterli.The Finite Ridgelet Transform for Image Representation.IEEE Trans on Image Processing.2003.Vol.12(No.1):16-27 P
[63]http://cobweb.ecn.purdue.edu/～biehl/MultiSpec/
[64]http://stat.stanford.edu/～donoho/reports.html
[2]童庆禧,张兵,郑兰芬.高光谱遥感——原理、技术与应用.北京:高等教育出版社,2006:166-218页,293-316页
[3]浦瑞良,宫鹏.高光谱遥感及其应用.北京:高等教育出版社,2000:1-45页,123-228页
[4]徐青,张艳,耿则勋,邢帅,谭兵.遥感影像融合与分辨率增强技术.北京:科学出版社,2007:68-101页
[5]边肇祺,张学工.模式识别.北京:清华大学出版社,1999
[6]刘春红.超光谱遥感图像降维及分类方法研究.哈尔工程大学博士论文,2005:1-20页
[7]马建文,李启青,哈斯巴干,戴芹.遥感数据智能处理方法与程序设计.北京:科学出版社,2005
[8]Lee Sanghoon,Crawford Melba M.Hierarchical Clustering Approach for Unsupervised Image Classification of Hyperspectral Data.International Geoscience and Remote Sensing Symposium(IGARSS).2004,2:941-944 P
[9]Jimenez Luis O,Rivera-Medina Jorge L.Integration of spatial and Spectral Information by means of Unsupervised Extraction and Classification for Homogenous Objects Applied to Multispectral Data.IEEE Transactions on Geosciences and Remote Sensing.2005,43(4):844-851 P
[10]Du Qian.Independent Component Analysis to Hyperspectral Image Classification.Proceedings of SPIE.-The International Society for Optical Engineering.Imaging Spectrometry X.2004,5546:366-373 P
[11]Shah Chintan A,Arora Manoj K,Varshey Pramod K.Unsupervised Classification of Hyperspectral data:An ICA Mixture Model based Approach. International Journal of Remote Sensing. 2004,25 (2) :481-487 P
[12] Dudar M.M, Landgrebe D.A. Toward an Optimal supervised Classifier for the Analysis of Hyperspectral data. IEEE Transactions on Geoscience and Remote Sensing. 2004, 42 (1) : 271-277 P
[13] Rand Robert S. Spectral/Spatial Annealing of Hyperspctral Imagery Initialized by a Supervised Classification Method. Proceeding of SPIE—the International Society for Optical Engineering. 2002, 4725:242-253 P
[14] Landgrebe David, Murat Dunder M. A Model-based Mixture-supervised Classification Approach in Hyperspectral Data Analysis. IEEE Transactions on Geoscience and Remote Sensing. 2002, 40 (12) :2692-2699 P
[15] Chang Yang-Lang, Han Chin-Chuan. Greedy Modular Eigenspaces and Positive Boolean Function for Supervised Hyperspectral Image Classification. Optical Engineering. 2003, 42 (9): 2576-2587 P
[16] Rosario Dalton. A Semiparamrtric Approach Using Discriminant Metric SAM (Spectral Angle Mapper). Proceeding of SPIE—The International Society for Optical Engineering. 2004, 5426:58-66 P
[17] Goetz Alexandar F.H, Vane Gregg, et al. Image Spectrometry for Earth Remote Sensing. Science. 1985, 228 (4704) : 1147-1153 P
[18] Schmidt K.S, Skidmore A.K, et al. Mapping Coastal Vegetation Using an Expert System and Hyperspectral Imagery. Photogram metric Engineering and Remote Sensing. 2004, 70 (6) : 703-715 P
[19] Gustavo Camps-Vails, Lorenzo Bruzzone. Kernel-Based Methods for Hyperspectral Image Classification. IEEE Transactions on Geoscience and Remote Sensing. Vol.43, No.6, 2005, 7:1351-1362 P
[20] Jimenez L O, et al. Classification of Hyperdimensional Data Based on Feature and Decision Fusion Approaches Using Projection Pursuit,Majority Voting, and Neural Networks. IEEE on G.R.S. 1999, 37(3):1360-1366 P
[21] Jon Atli Benediktsson, Ioannis Kanellopoulos. Classification of Multisource and Hyperspectral Data Based on Decision Fusion. IEEE Trans on Geoscience and Remote Sensing. 1999, 37 (3) : 1367-1377 P
[22] Baofeng Guo, Steve Gunn, Bob Damper, James Nelson. Hyperspectral Image Fusion using Spectrally Weighted Kernels. 7th International conference on Information Fusion. 2005:402-408 P
[23] Anil Cheriyadat, Lori Mann Bruce, et al. Decision level fusion with best-Bases for Hyperspectral Classification. IEEE Trans on Geo science and Remote Sensing. 2004:399-406 P
[24] Jorge Nunez, et al. Multiresolusion-Based Image Fusion with Additive Decomposition. IEEE Trans on Geo science and Remote Sensing, 1999,37 (3): 1204-1211 P
[25] Chunhong Liu, Chunhui Zhao, WAnhai Chen. Hyperspectral Image Classification by Second Generation Wavelet Based on Adaptive Band Selection. Proceedings of the IEEE International Conference on Mechatronics & Automation. 2005:1175-1179 P
[26] Terry A Wilson, Steven K Rogers, Matthew Kabrisky. Perpetual-Based Image Fusion for Hyperspectral Data. IEEE Trans on Geo science and Remote Sensing, 1997, 35 (4) : 1007-1017 P
[27] Jon Atli Benediktsson, Jon Aevar Palmason, et al. Decision Level Fusion in Classification of Hyperspectral Data from Urban Areas.IEEE Trans on Geo science and Remote Sensing. 2004:73-76 P
[28] P.C. Smits. Comparsion of some Feature Subset Selection Methods for Use in Remote Sensing Image Analysis. IEEE Trans on Geo science and Remote Sensing. 2001:530-533 P
[29] Bor-Chen Kuo, et al. Hyperspectral Data Calssification Using Classifier Overproduction and Fusion Strategies. IEEE Trans, 2004:2937-2940 P
[30] Y. Allard, A. Jouan. Interpretation of high-resolution optical imagery with evidential fusion of spectral information and object attributes.9th International Conference on Information Fusion,2006:1-7 P
[31]Boriana L.Milenova,Marcos M.Campos.Mining High-Dimensional Data for Information Fusion:A Database-centric Approach.7th International Conference on Information Fusion.2005:638-645 P
[32]Junping Zhang,Ye Zhang.Fusion Classification of Hyper-spectral Image Based on Adaptive Subspace Decomposition.International Conference on Image Processing,Vol.3,2000:10-13 P
[33]张钧萍,张晔.基于多特征多分辨率融合的高管够图像分类.红外与毫米波学报,2004,Vol.23.No.5:345-348页
[34]Qiang Wang,Yi Shen.Performances Evaluation of Image Fusion Techniques Based on Nonlinear Correlation Measurement.Instrumentation and Measurement Technology Conference.2004:472-475 P
[35]Liang Sun,Chongzhao Han,Ming Lei.Knowledge discovery-based multiple classifier fusion:a Generalized Rough Set Method.9th International Conference on Information Fusion.2006:1-8 P
[36]孙亮,韩崇昭,康欣.关于集值特征属性的知识发现方法研究.Proceedings of the 5th World Congress On Intelligent Control and Automation.2004:3051-3054 P
[37]Joseph C Harsanyi,Chein-IChang.Hyperspectral Image Classification and Dimensionality Reduction:An Orthogonal Subspace Projection Approach.IEEE Trans on G.R.S,1994,32(4):779-785 P
[38]孙延奎.小波分析与应用.北京:机械工业出版社,2005
[39]赵英时.遥感应用分析原理与方法.北京:科学出版社,2003:190-208页
[40]David Landgrebe.Multispectral Data Analysis:A Signal Theory Perspective.http://cobweb.ecn.purdue.edu/%TEbiehl/MultiSpec/
[41]阎平凡,张长水.人工神经网络与模拟进化计算.北京:清华大学出版社,2005
[42]孙即祥.现代模式识别.长沙:国防科技大学出版社,2002
[43]张代远.神经网络新理论与方法.北京:清华大学出版社,2006:51-113页
[44]Landgrebe D.Information Extraction Principles and Methods for Multispectral and Hyperspectral Image Data,Chapter 1 of Information Processing for Remote Sensing,edited by C.H.Chen,published by the World Scientific Publishing Co.Inc.,1060 Main Street,River Edge,NJ07661,USA,2000
[45]Jimenez Luis O,Rivera-Medina Jorge L.Integration of spatial and spectral information by means of unsupervised extraction and classification for homogenous objects applied to multispectral and hyperspectral data.IEEE Transactions on Geoscience and Remote Sensing.2005,43(4):844-851 P
[46]Hughes G..F.On the mean accuracy of statistical pattern recognizers.IEEE Trans.On Information Theory,Jan 1968,Vol.IT-14,No.1
[47]李弼程,魏俊,彭天强.遥感影象融合效果的客观分析与评价.计算机工程与科学,2004,Vol.26,No.1:42-46页
[48]崔岩梅,倪国强.利用统计特性进行图像融合效果分析及评价.北京理工大学学报,2000,20(1):102-146页
[49]E.J.Candes.Ridgelets:Theory and Application.The Department of Statistics and the Committee on Graduate Studies of Stanford University.Stanford University.1998
[50]黄建平,王卫卫.基于二进脊波变换的图像去噪.计算机工程与应用,2006,05:69-79页
[51]Starck J.L,Candes E.J.,Donoho D.L.The curvelet transform for image denoising.IEEE Trans.IP.2002,11(6):670-684 P
[52]吴昊.高光谱遥感图像数据分类技术研究.国防科学技术大学博士学位论文,2004
[53]Candes E.J.,Donoho D.L.Curvelets,Multiresolution Representation,and Scaling Laws.SPIE Wavelet Application in Signal and Image Processing ⅤⅢ,2002,Vol.4119
[54]刘春红,赵春晖.基于提升算法的超谱遥感图像融合分类研究.哈尔滨工程大学学报,2004,Vol.25(No.6):794-798页
[55]Tao Chen,Junping Zhang.Remote Sensing Image Fusion Based on Ridgelet Transform.Geoscience and Remote Sensing Symposium,2005:1150-1153 P
[56]A.Averbuch,R.R.Coifman,D.L.Donoho.Fast Slant Stack:A notion of Radon Transform for Data in a Cartesian Grid which is Rapidly Computible,Algebraically Exact,Geometrically Faithful and Invertible.http://stat.stanford.edu/～donoho/reports
[57]Donoho D.L,Flesia A.G.Digital Ridgelet Transform based on True Ridge Functions.http://stat.stanford.edu/～donoho/reports
[58]白键,冯象初.基于局部对偶框架的数字脊波重构.中国科学(E辑)信息科学,2005,35(10):1072-1082页
[59]张代远.样条权函数神经网络的一种新算法.系统工程与电子技术,2006(9):1434-1437页
[60]张钧萍,张晔,周显廷.基于信息融合的超谱图像分类方法研究.哈尔滨工业大学学报,2002,34(4):464-468页
[61]Zhang Ye.Adaptive subspace decomposition for hyperspectral data dimensionality reduction.ICIP99 Proceedings.1999:326-329 P
[62]Minh N.Do,Martin Vetterli.The Finite Ridgelet Transform for Image Representation.IEEE Trans on Image Processing.2003.Vol.12(No.1):16-27 P
[63]http://cobweb.ecn.purdue.edu/～biehl/MultiSpec/
[64]http://stat.stanford.edu/～donoho/reports.html