肖奥,赵文吉,胡德勇,刘连刚,李家存

• 1:首都师范大学资源环境与旅游学院/三维信息获取与应用教育部重点实验室/资源环境与地理信息系统北京市重点实验室
• 2:北京市地质研究所

摘要(Abstract)：

土地利用/覆被专题信息的快速、高效、准确提取是遥感图像处理研究的重要方向。传统的遥感分类方法常依靠像元的光谱值,未充分利用影像的空间信息。本文将面向对象影像分割和支持向量机方法相结合,复合光谱和纹理信息,建立了Object-SVM分类模型,并与面向对象的模糊函数和基于像元的SVM方法相比较,探寻区域尺度土地利用/覆被信息提取方法。结果显示,Object-SVM模型有效地提高了遥感图像的分类精度和分类效率,对于区域尺度影像的快速、准确、客观的信息提取具有实际意义。

关键词(KeyWords)： 面向对象;影像分割;支持向量机;光谱;纹理

Abstract:

Keywords:

基金项目(Foundation): 国家科技支撑计划课题“环北京区域地表环境遥感动态监测与评价技术研究”(2007BAH15B02),国家科技支撑计划重点(2006BAC08B02);; 空间数据挖掘与信息共享教育部重点实验室开放基金“基于对象和SVM的遥感图像分类及精度分析”

作者(Author): 肖奥,赵文吉,胡德勇,刘连刚,李家存

DOI: 10.16251/j.cnki.1009-2307.2010.05.075

参考文献(References)：

• [1]HUANG HP,WUB F,LI MM.Detecting Urban Vege-tation Efficiently with High Resolution Remote SensingData[J].International Journal of Remote Sensing,2004,8(1).
• [2]赵书河,冯学智,都金康,林广发.基于支持向量机的SPIN-2影像与SPOT-4多光谱影像融合研究[J].遥感学报,2003,9(1).
• [3]S V STEHMAN,TLSOHL,TR LOVELAND.An Evalua-tion of Sampling Strategies to Improve Precision of Estimatesof Gross Change in Land Use and Land Cover[J].Inter-national Journal of Remote Sensing,2005,26(20).
• [4]ZHOU W,TROY A.An Object-oriented Approach forAnalysing and Characterizing Urban Landscape at theParcel Level[J].International Journal of RemoteSensing,2008,29(11).
• [5]FRIEKE M B VAN COILLIE,LIEVEN PC VERBEKE,ROBERTR DE WULF.Feature Selection by Genetic Al-gorithms in Object-based Classification of IKONOS Im-agery for Forest Mapping in Flanders,Belgium[J].Remote Sensing of Environment,2007,110.
• [6]KARATHANASSI V,TOPOUZELIS K,PAVLAKIS.AnObject-oriented Methodology to Detect Oil Spills[J].In-ternational Journal of Remote Sensing,2006,27(23).
• [7]CHEN Y,SHI P,FUNG T.Object-oriented Classificationfor Urban Land Cover Mapping with ASTER Imagery[J].International Journal of Remote Sensing,2007,28(20).
• [8]LUCIEER V L.Object-oriented Classification of Side ScanSonar Data for Mapping Benthic Marine Habitats[J].In-ternational Journal of Remote Sensing,2008,29(3).
• [9]WEI SU,JING LI,YUNHAO CHEN.Textural and LocalSpatial Statistics for the Object-oriented Classification ofUrban Areas Using High Resolution Imagery[J].In-ternational Journal of Remote Sensing,2008,29(11).
• [10]李国正,王猛,曾华军译.支持向量机导论[M].北京:电子工业出版社,2006.
• [11]CHENGQUAN HUANG,KUAN SONG,SUNGHEEKIM.Use of a Dark Object Concept and Support VectorMachines to Automate Forest[J].Remote Sensing ofEnvironment,2008,112.
• [12]DEMIR B,ERTRK S.Phase Correlation Based Redun-dancy Removal in Feature Weighting Band Selection forHyperspectral Images[J].International Journal of Re-mote Sensing,2008,29(6).
• [13]HUANG H,GONG P,Clinton N.Reduction of Atmos-pheric and Topographic Effect on Landsat TM Data forForest Classification[J].International Journal of Re-mote Sensing,2008,29(19).
• [14]HUGO CARRAO,PAULO GONCALVES,MáRIOCAETANO.Contribution of Multispectral and Multitem-poral Information from MODIS[J].Remote Sensing ofEnvironment,2008,112.
• [15]TANG S,CHENC,ZHANH.Determination of Ocean Pri-mary Productivity Using Support Vector Machines[J].In-ternational Journal of Remote Sensing,2008,29(21).
• [16]章毓晋.图像分割[M].北京:北京.科学出版社,2001.
• [17]CHANG C C,LIN C J.LIBSVM:A Library for SupportVector Machines[EB/OL].(2006-08-27).http://www.csie.ntu.edu.tw/~cjlin/libsvm.