文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > 自然科学 > 物理 > 面向对象的高分辨率遥感影像分割分类评价指标
侵权投诉

面向对象的高分辨率遥感影像分割分类评价指标

4期吴波等:面向对象的高分辨率遥感影像分割分类评价指标573

[9]FoodyGM.Onthecompensationforchanceagreementin

imageclassificationaccuracyassessment[J].Photogram-metricEngineeringandRemoteSensing,1992,58(10):1459-1460.

[10]PerselloC,BruzzoneL.Anovelprotocolforaccuracyas-sessmentinclassificationofveryhighresolutionimages[J].IEEETransactionsonGeoscienceandRemoteSens-ing,2009,48(3):1232-1244.

[11]HuangQ,DomB.Quantitativemethodsofevaluatingim-agesegmentation[C].Proc.Conf.onImageProcessing(ICIP’95),1995(3):53-56.

[12]章毓晋.图像工程(第2版)[M].北京:清华大学出版社,

2007.

[13]ComanicisD,MeerP.Meanshift:Arobustapplicationto-wardfeaturespaceanalysis[J].IEEETransactionsonPat-ternAnalysisandMachineIntelligence,2002,24(5):603-619.

[14]沈占锋,骆剑承,胡晓东.高分辨率遥感影像多尺度均值

漂移分割算法研究[J].武汉大学学报·信息科学版,2010,35(3):313-316.

[15]薄树奎,韩新超,丁琳.面向对象影像分类中分割参数的

选择[J].武汉大学学报·信息科学版,2009,34(5):514-517.

QuantitativelyEvaluatingIndexesforObject-basedSegmentationofHigh

SpatialResolutionImage

WUBo*,LINShanshanandZHOUGuijun

(KeyLaboratoryofSpatialDataMiningandInformationSharingofMinistryofEducation,SpatialInformation

ResearchCenterofFujianProvince,FuzhouUniversity,Fuzhou350002,China)

Abstract:TraditionalclassificationaccuracyassessmentsintermsofoverallaccuracyorKappacoefficientbasedonisolatingpixelsstatistic,cannotcapturethegeometricalpropertiesofthesegmentedimageobjects,andthereforedonotprovideanaccurateevaluationinobjectbasedsegmentationorclassification.Consideringtheimportanceofborderingpixelsandgeometricshapeinobjectsegmentation,itisthusofourinterestingtodesignsomemeasuringindexesintroducedtheborderpixelsandgeometricstructureinformationtoevaluatetheob-ject-basedresultantimagesegmentationorclassification.ThispaperimprovedfiveindexesoriginallyproposedbyPerselloandBruzzonetoevaluatetheresultantsegmentationofhighspatialresolutionimage.Realizingthattheoriginalindicescannotwellcapturethegapamongdifferentcategories,resultingindifficultiesindiscriminat-ingtheassessmentresultsofdifferentcategories,weusenormalizedtechniquesbasedongeometricshapetoovercomeit.Theindicesdependonthegeometryfeaturesofeachobjectofthethematicmapincludingoverseg-mentation,undersegmentation,edgelocation,fragmentationerrorandshapeerror.Moreover,werealizedapro-totypesystemcontainedtheaforementionedevaluatedindexesbasedonIDLplatformtosupporttheob-ject-basedimageprocessingandanalysis.Tovalidatetheseindexes,asubsetQuickBirdimagelocatedonFu-zhouwasimplementedandtheresultsoftheMeanshiftsegmentedalgorithmdemonstratethattheproposedin-dexescanprovidebetteraccuracyevaluationofeachlandcoverclass,andcanmakeusersmoreeffectivelychoosethebestclassificationmap.Moreover,ourexperimentsalsodemonstratethat,comparedwithOAandKappacoefficient,theproposedindexeshaveadvantagesoncharacterizingthedetailedgroundmaterialsandarehelpfulinaidingtheoptimalparametersselectionfortheMeanshiftsegmentedalgorithm.

Keywords:high-resolutionremotesensingimage;segmentation/classificationevaluation;object-based;geomet-ricalstructure

*Correspondingauthor:WUBo,E-mail:wavelet778@sohu.com

上一页第7页

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email