411国家重点基础研究发展计划(2009CB219302)资助收稿日期: 2011-06-17; 修回日期: 2011-08-22; 网络出版日期: 2012-02-24网络出版地址: /kcms/detail/11.2442.N.20120224.1047.011.html北京大学学报(自然科学版), 第48卷, 第3期, 2012年5月Acta Scientiarum Naturalium Universitatis Pekinensis, Vol. 48, No. 3 (May 2012)运用EO-1 Hyperion 数据和单类支持向量机方法提取岩性信息张西雅 徐海卿 李培军†北京大学地球与空间科学学院遥感与地理信息系统研究所, 北京 100871; †通信作者, E-mail: pjli@摘要 将扩展的单类支持向量机方法运用到高光谱岩性识别中, 并分析和评价该方法的性能。

利用单类支持向量机分别提取各个感兴趣的岩性类别, 对于被识别为多个岩性类别的像元, 根据该像元与每个单类支持向量机所确定的分类超平面的距离来确定属于哪一类别, 这样, 利用扩展的单类支持向量机来可提取多个感兴趣的岩性类别。

将该方法运用到新疆准噶尔地区的EO-1 Hyperion 高光谱数据岩性分类中, 并与传统的光谱角制图方法进行比较。

结果表明, 扩展的单类支持向量机方法的岩性分类精度显著高于光谱角制图方法, 是一种新的可用于高光谱数据的岩性分类方法。

关键词 高光谱; 单类支持向量机; 光谱角制图; 岩性分类 中图分类号 P627Lithologic Mapping Using EO-1 Hyperion Data and Extended OCSVMZHANG Xiya, XU Haiqing, LI Peijun Key words hyperspectral data; OCSVM; SAM; lithologic classification参考文献†Institute of Remote Sensing and Geographic Information System, Peking University, Beijing 100871;† Corresponding author, E-mail: pjli@Abstract An extended one-class support vector machine (OCSVM) was applied to lithologic mapping from the EO-1 Hyperion hyperspectral data, and it has been evaluated in terms of classification accuracy. First OCSVM was separately used to extract each lithologic unit of interest. The pixel which was classified to different classes simultaneously was then assigned as the class with smallest distance to the hyperplane. In this way, the extended OCSVM can be used for extracting several lithologic units of interest. The extended OCSVM method was used in lithologic classification from the EO-1 Hyperion hyperspectral data in Junggar area, Xinjiang and compared with the spectral angle mapper (SAM) method. The results showed that the extended OCSVM method outperformed the SAM method in lithologic classification. The extended OCSVM is a useful and effective method for lithologic classification from hyperspectral remote sensing data.[1] Chica-Olmo M, Arbarca-Hernandez F. Computinggeostatistical image texture for remotely sensed data classification. Computers & Geosciences, 2000, 26(4): 373–383[2] Schetselaar E M, Chung J F, Kim K E. Integrationof Landsat TM, Gamma-ray, magnetic, and field data to discriminate lithological units in vegetated granite-gneiss terrain. Remote Sensing of Environment, 2000, 71: 89–105[3] Rowan L C, Mars J C. Lithologic mapping in themountain pass, California area using advanced spaceborne thermal emission and reflection北京大学学报(自然科学版) 第48卷412 radiometer (ASTER) data. Remote Sensing of Environment, 2003, 84: 350–366[4]Gomez C, Delacourt C, Allemand P, et al. UsingASTER remote sensing data set for geologicalmapping in Namibia. Physics and Chemistry of theEarth, 2005, 30: 97–108[5]Ninomiya Y, Fu B H, Cudahy T J. Detectinglithology with advanced spaceborne thermalemission and reflection radiometer (ASTER)multispectral thermal infrared “radiance at sensor”data. Remote Sensing of Environment, 2005, 99:127–139[6]Fu B H, Zheng G D, Yoshiki N, et al. Mappinghydrocarbon-induced mineralogical alterations inthe northern Tianshan using ASTER multispectraldata. Terra Nova, 2007, 19: 225–231[7]李培军, 龙玄耀, 刘立. 基于ASTER数据的蛇绿岩组分识别: 以德尔尼矿区为例. 岩石学报, 2007,23(5): 1175–1180[8]Crosta A P, Sabine C, Taranik J V. Hydrothermalalteration mapping at Bodie, California, usingAVIRIS hyperspectral data. Remote Sensing ofEnvironment, 1998, 65(3): 309–319[9]Rowan L C, Simpson C J, Mars J C. Hyperspectralanalysis of the ultramafic complex and adjacentlithologies at Mordor. Remote Sensing ofEnvironment, 2004, 91(30): 419–431[10]Sanchez-Hernandez C, Boyd D S, Foody G M.One-class classification for mapping a specificland-cover class: SVDD classification of Fenland.IEEE Transactions on Geoscience and RemoteSensing, 2007, 45(4): 1061–1073[11]Kruse F A, Lefkoff A B, Boardman J B, et al. Thespectral image processing system (SIPS) —interactive visualization and analysis of imagingspectrometer data. Remote Sensing of Environ-ment, 1993, 44: 145–163[12]Guo Q, Kelly M, Graham C. Support vectormachines for prediction distribution of sudden oakdeath in California. Ecological modeling, 2005,128(1): 75–90[13]Muñoz-Marí J, Bruzzone L, Camps-Valls G. Asupport vector domain description approach tosupervised classification of remote sensing images.IEEE Transactions on Geoscience and RemoteSensing, 2007, 45(8): 2683–2792 [14]Li P, Xu H. Land-cover change detection usingone-class support vector machine. PhotogrammetricEngineering and Remote Sensing, 2010, 76(3):255–263[15]Li P, Xu H,Guo J. Urban building damage detectionfrom very high resolution imagery using one-classSVM and spatial features. International Journal ofRemote Sensing, 2010, 31(13): 3393–3409[16]Barry P. EO-1/hyperion science data user’s guide.Redondo Beach, CA: TRW Space, Defense &Inform Syst, 2001[17]Beck R. EO-1 User guide v.2.3. (2003)[2011]. http://&[18]新疆维吾尔自治区地质矿产局. 巴里坤湖地质图(1:20万). 北京: 地质出版社, 1961[19]新疆维吾尔自治区地质矿产局. 东泉幅地质图(1:20万). 北京: 地质出版社, 1981[20]Matthew M W, Adler-Golden S M, Berk A, et al.Status of atmospheric correction using aMODTRAN4-based algorithm. Proc SPIE, 2000,4049: 199–207[21]Matthew M W, Adler-Golden S M, Berk A, et al.Atmospheric correction of spectral imagery:evaluation of the FLAASH algorithm with AVIRISdata. Proc SPIE, 2003, 5093: 474–482[22]Adler-Golden S M, Acharya P K, Berk A, et al.Remote bathymetry of the littoral zone fromAVIRIS, LASH, and quickbird imagery. IEEETransactions on Geoscience and Remote Sensing,2005, 43(2): 337–347[23]Hughes G F. On the mean accuracy of statisticalpattern recognizers. IEEE Transactions on Information Theory, 1968, 14(1): 55–63[24]Li P, Moon W M. Land cover classification usingMODIS/ASTER airborne simulator (MASTER)data and NDVI: a case study of the Kochang area,Korea. Canadian Journal of Remote Sensing, 2004,30(2): 123–136[25]Zhang X, Pazner M. Preprocessin, featureextraction and lithologic mapping using EO-1hyperion data. Journal of Image and Graphics,2007, 12(6): 981–990[26]Green A A, Berman M, Switzer P, et al. Atransformation for ordering multispectral data in第3期张西雅等: 运用EO-1 Hyperion数据和单类支持向量机方法提取岩性信息terms of image quality with implications for noiseremoval. IEEE Transactions on Geoscience andRemote Sensing, 1988, 26(1): 65–74[27]Tax D M J. One-class classification [D]. Delft, TheNetherlands: Delft University of Technology, 2001[28]Scholköpf B, Platt J C, Shawe-Taylor J, et al.Estimating the support of a highdimensionaldistribution. Microsoft Research Technique Report,MSR-TR-99-87. 1999, 11[29]Tax D M J, Duin R P W. Support vector domaindescription. Pattern Recognition Letters, 1999, 20:1191–1199[30]Belousov A I, Verzakov S A, von Frese J. A flexibleclassification approach with optimal generalizationperformance: support vector machines. Chemo-metrics and Intelligent Laboratory Systems, 2002,64: 15–25[31]Duin R, Tax D. Experiments with classifiercombining rules // Proc Int Workshop MCS.Cagliari, Italy, 2000, 1875: 16–29[32]Tax D M J, Duin R P W. Combining one-classclassifiers // Proc Int Workshop MCS. Cambridge,2001, 2096: 299–308[33]Lee D, Lee J. Domain described support vectorclassifier for multi-classification problems. PatternRecognit, 2007, 40(1): 41–51[34]Unnthorsson R, Runarsson T P, Jonsson M T.Model selection in one-class ν-SVMs using RBFkernels // Proceedings of 16th International Con-gress and Exhibition on Condition Monitoring andDiagnostic Engineering Management (COMADEM2003). Vexjo University, Sweden, 2003: 27–29[35]时丕龙. 基于ASTER多光谱遥感的岩性信息及烃类渗漏异常信息提取与识别方法研究[D]. 北京: 中国科学院地质与地球物理研究所, 2011413运用EO-1 Hyperion数据和单类支持向量机方法提取岩性信息作者：张西雅， 徐海卿， 李培军作者单位：北京大学地球与空间科学学院遥感与地理信息系统研究所,北京100871刊名：北京大学学报（自然科学版）英文刊名：Acta Scientiarum Naturalium Universitatis Pekinensis年，卷(期)：2012(3)1.Chica-Olmo M.Arbarca-Hernandez F Computing geostatistical image texture for remotely sensed data classification2000(04)2.Schetselaar E M.Chung J F.Kim K E Integration of Landsat TM,Gamma-ray,magnetic,and field data to discriminate lithological units in vegetated granite-gneiss terrain 20003.Rowan L C.Mars J C Lithologic mapping in the mountain pass,California area using advanced spaceborne thermal emission and reflection radiometer(ASTER)data 20034.Gomez C.Delacourt C.Allemand P Using ASTER remote sensing data set for geological mapping in Namibia 20055.Ninomiya Y.Fu B H.Cudahy T J Detecting lithology with advanced spaceborne thermal emission and reflectionradiometer(ASTER)multispectral thermal infrared"radiance at sensor”data 20056.Fu B H.Zheng G D.Yoshiki N Mapping hydrocarbon-induced mineralogical alterations in the northern Tianshan using ASTER multispectral data 20077.李培军.龙玄耀.刘立基于ASTER数据的蛇绿岩组分识别:以德尔尼矿区为例 2007(05)8.Crosta A P.Sabine C.Taranik J V Hydrothermal alteration mapping at Bodie,California,using AVIRIS hyperspectral data 1998(03)9.Rowan L C.Simpson C J.Mars J C Hyperspectral analysis of the ultramafic complex and adjacent lithologies at Mordor 2004(30)10.Sanchez-Hernandez C.Boyd D S.Foody G M One-class classification for mapping a specific land-cover class:SVDD classification of Fenland 2007(04)11.Kruse F A.Lefkoff A B.Boardman J B The spectral image processing system(SIPS)-interactive visualization and analysis of imaging spectrometer data 199312.Guo Q.Kelly M.Graham C Support vector machines for prediction distribution of sudden oak death in California2005(01)13.Mu?oz-Marí J.Bruzzone L.Camps-Valls G A support vector domain description approach to supervised classification of remote sensing images 2007(08)14.Li P.Xu H Land-cover change detection using one-class support vector machine 2010(03)15.Li P.Xu H.Guo J Urban building damage detection from very high resolution imagery using one-class SVM and spatial features 2010(13)16.Barry P EO-1/hyperion science data user’s guide 200117.Beck R EO-1 User guide v,2.3 200318.新疆维吾尔自治区地质矿产局巴里坤湖地质图(1:20万) 196119.新疆维吾尔自治区地质矿产局东泉幅地质图(1:20万) 198120.Matthew M W.Adler-Golden S M.Berk A Status of atmospheric correction using a MODTRAN4-based algorithm 200021.Matthew M W.Adler-Golden S M.Berk A Atmospheric correction of spectral imagery:evaluation of the FLAASH algorithm with AVIRIS data 200322.Adler-Golden S M.Acharya P K.Berk A Remote bathymetry of the littoral zone from AVIRIS,LASH,and quickbird imagery 2005(02)23.Hughes G F On the mean accuracy of statistical pattern recognizers 1968(01)24.Li P.Moon W M Land cover classification using MODIS/ASTER airborne simulator(MASTER)data and NDVI:a case study ofthe Kochang area,Korea 2004(02)25.Zhang X.Pazner M Preprocessin,feature extraction and lithologic mapping using EO-1 hyperion data 2007(06)26.Green A A.Berman M.Switzer P A transformation for ordering multispectral data in terms of image quality with implications for noise removal 1988(01)27.Tax D M J One-class classification 200128.Scholk?pf B.Platt J C.Shawe-Taylor J Estimating the support of a highdimensional distribution.Microsoft Research Technique Report,MSR-TR-99-87 199929.Tax D M J.Duin R P W Support vector domain description 199930.Belousov A I.Verzakov S A.von Frese J A flexible classification approach with optimal generalization performance:support vector machines 200231.Duin R.Tax D Experiments with classifier combining rules 200032.Tax D M J.Duin R P W Combining one-class classifiers 200133.Lee D.Lee J Domain described support vector classifier for multi-classification problems 2007(01)34.Unnthorsson R.Runarsson T P.Jonsson M T Model selection in one-class ν-SVMs using RBF kernels 200335.时丕龙基于 ASTER 多光谱遥感的岩性信息及烃类渗漏异常信息提取与识别方法研究 2011本文链接：/Periodical_bjdxxb201203013.aspx。