城市建设用地遥感信息提取方法研究

35420107ScienceofSurveyingandMappingVol35No4Jul:(1982),,,,Emai:lchefeng199@yahoocomcn:20090108:(30471391);(04B059);;(07005B),(,410004)本文利用CBERS02B影像,通过分析各个地类的光谱特征,发现了长沙市城市建设用地和其他背景地物的区别,并在此基础上选取了土壤调节植被指数(SAVI)归一化水体指数(NDWI)和比值居民地指数(RRI)作为三个指数波段,重新进行波段组合,从而减少了波段数据的冗余,最后采用最大似然法分类,提取出城市建设用地信息,其正确率达到856%CBERS02B影像;城市建筑用地;土壤调节植被指数TP79A10092307(2010)040097031,,,,,,[1],,[2],Gong[3]RiddV2I2S(22),3[4]MasekNDVI1973~1996[5][6][7]Rashed[8]Zhang,,[9]YongLandsatETM+[10]2007919,02BCCDWFI,02BHR,,,[11,12]CBERSLandsatTM,SPOT,SPOTCBERS,40%~50%,,200641CBERS[13]CBERS,,CBERS02B,,,221,,111!53∀~114!15∀27!51∀~28!41∀,,,55633km2,168!#172!200812CBERS02B236mHR195mCCD22CBERS02BCCDB1(045~052m)B2(052~059m)B3(063~069m)B4(077~089m)B5(051~073m)1B1,;B2,,,;B3,;B4,,1DNB14721342839144144449B23735279129913692373B35983312396345496393B455495897203486976106B543182664304736844475/31314417515328502B,,352120~170,DN1(2)12,DN::B3B4B1B5B2:B4B1B3B2B5:B3B1B5B2B4:B4B3B1B2B5:B3B4B5B1B2:B2B3B4B5,B3B4,DN20,343DN4DN,,(),B4B3B4/B3,(),(SAVI),11,,,B2B4,(NDWI),,232311973Rouse(NDVI),,,,,,,,NDVI,,1988HueteSAVI(soiladjustedvegetationindex),,,l,[14]SAVI=[(NIR-Red)(1+l)]/(NIR+Red+l)(1),NIR,Red,CBERS02B,43,l,0-1,∃0%,∃1%,l05,,(SAVI0,),(SAVI0,),SAVI(3b)3232McFeeters(NormalizedWaterIndex,NDWI),:NDWI=(Green-NIR)/(Green+NIR)(2),NIR,GreenCBERS02B,42(SAVI),3bSAVI,,(NDWI),4233(RatioResidentareaIndex,RRI)RVI(RatioVegetationIndex),[15],1,4,RRI=B1/B4,RRI(5),03438-25,1,RRI:0850807286,RRI(08508+07286)/2=078974NDWI5RRI234(SAVI),(NDWI),,,,SAVINDWIRRI,3,,,,984662(%)2744632085627153180853011995003,PanPan236m,195m,PanPan500,Geolink,,856%4SAVINDWIRRI,,,,,,,,,,,:&CBERS02B,,,;∋,,,,,,[1],,,[M](D),2002,32(12)[2],,,/[M]:,2000[3]GongP,HowarthPJTheuseofstructuralinformationforimprovinglandcoverclassificationaccuraciesattheruralurbanfringe[J]PhotogrammetricEngineeringandRemoteSensing,1990,56(1)[4]RiddMKExploringaVIS(Vegetationimpervioussurfacesoil)modelforurbanecosystemanalysisthroughremotesensing:comparativeanatomyforcities[J]InternationalJournalofRemoteSensing,1995,16(12)[5]MasekJG,LindsayFE,GowardSNDynamicsofurbangrowthinWashingtonDCmetropolitanarea,19731996,fromLandsatobservations[J]InternationalJournalofRemoteSensing,2000,(18)[6],TM[J],2000,4(2)[7][J],2002,17(3)[8]RashedT,WeeksJR,GadallaMS,etalRevealingtheAnatomyofcitiesthroughspectralmixtureanalysisofmultispectralsatellitemiagery:acasestudyoftheGreaterCairo,Egypt[J]GeocartoInternationa,l2001,16(4)[9]ZhangQ,WangJ,PengX,etalUrbanbuiltuplandchangedetectionwithroaddensityandspectralinformationfrommultitemporalLandsatTMdata[J]InternationalJournalofRemoteSensing,2002,23(15)[10]YangL,HuangC,HomerCG,etalAnapproachformappinglargeareaimpervioussurfaces:synergisticuseofLandsatETM+andhighspatialresolutionimagery[J].CanadianJournalofRemoteSensing,2003,29(2)[11][Z]2007,1:23[12],[J],2003,[13],[C]//,2003:4147[14]HueteARAsoiladjustedvegetationindex(SAVI)[J]RemoteSensingofEnvironment,1988,25(3)[15],,,[J](),2006:118121RemotesensinginformationextractionofurbanbuiltuplandAbstract:BasedontheCBERS02Bimagedata,thispaperanalyzedtheSpectralCharacteristicsoflandusetypes,andfoundthedifferencebetweentheurbanbuiltuplandandothersThestudyselectedthreeindices,ieSoilAdjustedVegetationIndex(SAVI),normalizeddifferencewaterindex(NDWI)andRatioResidentareaIndex(RRI),torepresentthreemajorindexbandsgeneratedfromtheoriginalmultispectralbandsanddramaticallydecreasedbandcorrelationUsingMaximumLikelihoodClassification(MLC)toretrieveurbanbuiltupareasintheregionofChangsha,theoverallaccuracyofthemwere856%Keywords:CBERS02Bimage;urbanbuiltupland;soiladjustedvegetationindexCHEFeng,LINHui(ResearchCentreofRemoteSensingandInformationEngineering,CentralSouthUniversityofForestry&Technology,Changsha410004,China)(108)AnidentificationmethodofroadjunctionsbasedonperceptualgroupingandstructuralcharacteristicsAbstract:Inordertoidentifyroadjunctionsfromvectormapefficientlyandprecisely,thispaperfirstproposedanidentificationmethodbasedonperceptualgroupingmethodandstructuralcharacteristicsoftheroadjunctionsTheexperimentscarriedouttovalidatethealgorithmwerethenintroduced,whichidentifiedtwotypicalareasofsimpleandcomplicatedroadnetworkrespectivelyTheresultsshowedthatthealgorithmcouldidentifyroadjunctionsefficientlyandpreciselywithstabilityandspeedKeywords:roadjunctions;identification;perceptualgrouping;structuralcharacteristicsMENGYanzi&∋,XUZhu&,LIUGuoxiang&,CAIGuolin&(&DeptofSurveyingEngineering,SouthwestJiaotongUniversity,Chengdu610031,China;∋SichuanRemoteSensingGeomaticsInstitute,Chengdu610100,China)99