4 : TM TM 1 5 TM DEM [5-6] 3 ; [7-9] / [5] [10] 3.1 TM :1 / ;2 DEM (ecognition) (1) / / / 2 ; [11] 2.1 Shannon ~ ~ TM 5TM

Size: px

Start display at page:

Download "4 : TM TM 1 5 TM DEM [5-6] 3 ; [7-9] / [5] [10] 3.1 TM :1 / ;2 DEM (ecognition) (1) / / / 2 ; [11] 2.1 Shannon ~ ~ TM 5TM"

督黄
5 years ago
Views:

1 RemoteSensingInformation Vol.28No.4 Aug.2013 朱海涛 12 张霞 1 王树东 1 王晋年 1 孙艳丽 12 ( ; ) : % : ; / ; ; doi: /j.issn :TP79 :A : (2013) Object-orientedandDecisionTreeClassificationfor RemoteSensingDatainSemi-aridAreas ZHU Hai-tao 12 ZHANG Xia 1 WANGShu-dong 1 WANGJin-nian 1 SUN Yan-li 12 (1.Instituteof RemoteSensing ApplicationsChinese Academyof SciencesBeijing100101; 2.Universityof Chinese Academyof SciencesBeijing100049) Abstract:Thecomplexityoftopographyandlandcoverinthesemi-aridareasbringsagreatdealofuncertaintiestothere- motesensingclassification.anobject-orienteddecisiontreeclassificationbasedon multi-featuresisestablishedinthispaper. Thedecisiontreeclassificationalgorithm makesfuluseofspectralfeaturesterraincharacteristicsphenologicalcharacteristics aswelascorrelationbetweenpixelstoextractthegroundinformation.theresultsshowthatthismethodcouldimprovethe classificationaccuracyremarkablyespecialyinthesemi-aridareas.theoveralaccuracyreaches % whichmeetsthere- quirementsofestimatingthenon-pointsourcepolution. Keywords:semi-aridareas;landuse/coverclassification;objectoriented;decisiontree 1 [2] / ( ) km 2 [1] :1 [3-4] / ;2 : : : (1986~) : haitaozhu09@gmail.com 50

2 4 : TM TM 1 5 TM DEM [5-6] 3 ; [7-9] / [5] [10] 3.1 TM :1 / ;2 DEM (ecognition) (1) / / / 2 ; [11] 2.1 Shannon ~ ~ TM 5TM 9TM 5 ~11 Band Band mm 155~184 Band2 Band Band2 Band Band Band Band Band Band Band (2) 2.2 :1 LandsatTM5 30m /34 TM 6 ;2 DEM 30m;3 ENVI / FLASSH 51

3 2013 [12-13] ;Huete 1997 (EVI)EVI [13] ; 2005 (UPDM) VIUPD UPDM ( ) 1~5 ; [14-15] EVI NDVIVIUPD [16-18] MNDWI ; 5 Rouse 1974 ( ) [10] (NDVI) ND SSI VI NDVI SSI ( 1(b)) 9 MNDWI( 2) MNDWI 2005 (MND- 9 5 WI) MNDWI (a)20105tm (b)20109tm ; ; RVI NDVI EVIVIUPD4 NDVI ( 1(a)) 5 9 RVI NDVI EVIVIUPD4 NDVI( 3) VIUPD 5 NDVI NDVI 9 VIUPD ( 4) NDVI VIUPD VIUPD VIUPD 5 (UPDM) CS( 5) 6 ~25 CS CS CS ( )

4 4 : 9 RVI NDVI EVIVIUPD4 EVI 9 EVI( 6) EVI EVI 9 EVI VIUPD 9 9 SSI ( 7) SSI SSI 9 SSI 5 CS 4 ; ; 6 EVI 2 MNDWI 7 SSI 3 NDVI MNDWI/TM5 ;2 9 NDVI ;3 9 SSI ; ;4 53

2013 ; ; 5 VIUPD ; 5 ; ; 5 UPDM CS ; 6 9 EVI 4 4.

5 2013 ; ; 5 VIUPD ; 5 ; ; 5 UPDM CS ; 6 9 EVI TM / / ( 9) 8 ( 59 ; 5 9 TM ) 9 / %

6 4 : Kappa ( 2) 80% % Kappa % ; 2 (Pixels) (%) (Pixels) (%) : % Kappa : % ; ; ; 80% 80% 99.77% 5 VIUPD UPDM / CSVIUPD CS VIUPD CS %Kappa / / 55

7 2013 : [1]. [J]. 2011(4):1-5. [2]. [J] (21): [3].SPOT5 [J] (3): [4] OUYANGZTZHANG M QXIEXetal.Acomparisonofpixel-basedandobject-orientedapproachestoVHRimage- ryformappingsaltmarshplants[j].ecologicalinformatics20116(2): [5]. / [J] (7): [6] MYINTS WGOBERPBRAZEL Aetal.Per-pixelvs.object-basedclassificationofurbanlandcoverextractionusing highspatialresolutionimagery[j].remotesensingofenvironment (15): [7]. [J] (2): [8] FRIEDL M ABRODLEY CE.Decisiontreeclassificationoflandcoverfromremotelysenseddata[J].RemoteSensing ofenvironment199761(3): [9] PAL MMATHERP M.Anassessmentoftheefectivenessofdecisiontreemethodsforlandcoverclassification[J].Re- motesensingofenvironment200386(4): [10]. LandsatTM [J] (22): [11]. [J] (2): [12]. [J] (4): [13]. [M]. : 2003: [14]. [D]. : [15] ZHANGLFFURUMISMURAMATSU Ketal.Anewvegetationindexbasedontheuniversalpaterndecomposi- tionmethod[j].internationaljournalofremotesensing200728(1): [16] MCFEETERSSK.Theuseofthenormalizeddiferencewaterindex(NDWI)inthedelineationofopenwaterfeatures [J].InternationalJournalofRemoteSensing199617(7): [17] XU H Q.Modificationofnormaliseddiferencewaterindex(NDWI)toenhanceopenwaterfeaturesinremotelysensed imagery[j].internationaljournalofremotesensing200627(14): [18]. (MNDWI) [J] (5): 櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂櫂 ( 23) : [1]. [J]. 2010(6): [2] HOOTSFRROEHRICH RL.SpacetrackreportNo.3-modelsforpropagationofNORADelementsets[R].Peter- son:aerospacedefensecommandunitedstatesairforce1980: [3]. [J] (6-3): [4]. [D]. : [5]. [D]. : [6] LIN W C.Dailyimagingschedulingofanearthobservationsatelite[J].IEEE TransactiononSystemsManandCyber- neticsparta:systemandhumans200501(35): [7] PINEDO M L.Scheduling:TheoryAlgorithmsandSystems(3rded)[M].Berlin:Spring2008. [8]. [M]. : 2011: [9]. [J] (3): [10]. [J]. ( )200040(9): [11]. [M]. : 2011:

DIS, 30 m TM (htp://glovis.usgs.gov/) ~2010 MODIS (htp://cdc.cma.gov.cn) Ⅰ 3 Ⅰ MODIS, :1 NDVI [6] NDVI NDVI, NDVI ; [7] 2 ; [

DIS, 30 m TM (htp://glovis.usgs.gov/) ~2010 MODIS (htp://cdc.cma.gov.cn) Ⅰ 3 Ⅰ MODIS, :1 NDVI [6] NDVI NDVI, NDVI ; [7] 2 ; [ 38 7 2013 7 GeomaticsandInformationScienceofWuhanUniversity Vol.38No.7 July2013 :1671-8860(2013)07-0789-05 :A MODIS 1,2,3 2 3 1 (1, 61,271018) (2 28,100830) (3 9,100094) : 250m MODIS 2000 TM MODIS 94.6%,,2010