空间抽样的理论方法与应用分析_以厦门岛问卷调查为例

12320106JOURNALOFGEO-INFORMATIONSCIENCEVol112,No13Jun1,2010:2010-01-25;:2010-04-15.:(KZCX2-YW-450);(40901243):(1985-),,,,E-mai:[e-mail]*:(1978-),,,,,GISE-mai:[e-mail])))高丽玲1,李新虎1*,王翠平1,邱全毅1,崔胜辉1,赵千钧2(1.,361021;2.,100101):,,,,,,,,,,,,:(1),,,,(2),,(3),,,,,,:;;;1,,,,,[1],[2],[3-8],[9-11],,,[12-19],,,,,,,,[20],,,2,GIS/0,,,,,118b04c04d24b26c46d)))1317km,1215km,128114km2,234km,,,,,199970km22007180km2,20012082008238,,,2008IKONOS3,,,,,GIS,/0,,,[21],()()3,.,1:1Fig11Flowchartofspatialsampling311,:,,,,,,,31111(1):0~200m,,,1,2,2,5b,50m,(1),(2)1Tab11Landformclassification(m)15b0~5025~8b50~20028~15b50~200215~25b50~200225~35b50~200235b50~2002Fig12MapofthephysiognomytypesinXiamenIsland3593:)))(2)2008,2062/km230720/km2,GIS,3,,[22],,,,5000/km2,7,3,,,筼筜湖鹭江道和轮渡附近密度接近为0,3Fig13DivisionofpopulationdensityinXiamenIsland(3),,,[23],,2000~5000/m2,,[24],5,4,筼筜湖和东南部的环岛路附近最高,在湖里区的北部,4Fig14ThebenchmarkpremiumofresidentiallandinXiamenCity(4)GIS,,,,,,,,02,,0,1;,,,23,,2,3,,2,5,筼筜湖,31112,,,,,,,6,3602010,2Tab12Hierarchicalcombinationoftheintegratedfactors1234501111112))2)22)))23))3)344)44)5))33)65)555:,21;/)05Fig15HierarchicalcombinationoftheintegratedfactorsinXiamenIsland312,,,:cost=c0+n*c1;,c0,,n,c1[25],100,,32,,,6Fig16StratifiedmapbyexpertinXiamenIsland,,,,7(a)7(b)7(c)7(d):(1)32,33,30,7(a),,,,筼筜湖和轮渡附近分布的点比(2),32,01001,7(b),,,,,2,;5,筼筜湖东渡和轮渡附近分布的样,80%,,,(3)32,01001,,,,7(c),,53613:))),50%,,,,,,,(4),0105,71405287,0105,7(d),,筼筜湖附近,也就是区域的各因素情况越,,,,,;,,,4,3622010,,,/0,:(1),,,,(2),,,,(3),,,,,,,,,,,,,,,,:[1].[D].,2007.[2],,.)))[J].,2009,16(9):110-114.[3],,,.[J].,2007,62(4):387-396.[4]YangJW.TransportationImplicationsofLandDevelop-mentinaTransitionalEconomy:EvidencefromHousingRelocationinBeijing[J].TransportationResearchRe-cord,2006:7-14.[5],.)))[J].,2004,23(2):227-241.[6].)))[J].,2007,26(5):1033-1042.[7],.[J].,2007(7):55-59.[8]WangDG,ChaiYW.TheJob-housingRelationshipandCommutinginBeijing,China:TheLegacyofDanwei[J].JournalofTransportGeography,2009,17:30-38.[9],.)))[J]..2004,23(2):227-241.[10],.[J].,2007(7):55-59.[11],.[J].,2009,28(5):1327-1330.[12].[J].,1994(6):74.[13],,,.GIS)))[J].,2008,24(6):30-34.[14],.GIS)))[J].,2005,22(2):274-276.[15],,,.[J].,2008,27(3):152-156.[16],.[J].,2002,12(5):545-547.[17].[J].,2007,22(1):5-10.[18].[J].,2003,18(5):24-29.[19],,,.[J].,2000,55(1):92-95.[20].[D].,2006.[21]WangJF,LiuJY,LiLF.SpatialSamplingDesignforMonitoringtheAreaofCultivatedLand[J].InternationalJournalofRemoteSensing,2002,23(2):263-284.[22],,,.[J].,2003,32(4):344-348.[23].[M].:,1989,22.[24].[J].,2006(6):37-40.[25],,,.[J].:,2004,34(10):975-982.3633:)))SurveySiteSelectionBasedontheSpatialSamplingTheory:ACaseStudyinXiamenIslandGAOLiling1,LIXinhu1,WANGCuiping1,QIUQuanyi1,CUIShenghui1,ZHAOQianjun2(1.KeyLaboratoryofUrbanEnvironmentandHealth,InstituteofUrbanEnvironment,CAS,Xiamen361021,China;2.InstituteofRemoteSensingApplications,CAS,Beijing100101,China)Abstract:Weoftenrunintotheproblemofhowtodistributesamplepointsintheprocessofsocialsurvey.Trad-itionalspatialsamplingtheoryisprimarilyusedinsingle-targe.tHowever,socialsurveyquestionnairessometimesaremult-iobjective.Andsomeobjectivesareattributedata,forexample,asurveyrelatedtopeople.sage,educationandothersocialinformation.Therefore,weshouldmakesomeimprovementsbasedontraditionalsamplingmethodstomeetcurrentrequirements.Dependingontheideaofmult-iobjectandspatialvariation,wemakeamult-iobjectivesurveyinthisstudy.InordertomeettheneedofsamplinginthequestionnairesurveyofXiamenIsland.ssettle-men,tweusethesandwichspatialsamplingmode,lbalancingthecostofsamplingandsamplingprecision.Ingener-a,lweadoptthreemethods:simplerandomsampling,spatialstratifiedsamplingandsandwichspatialsampling.A-ftercomparingofthosethreemethods,theresultsshowedtha:t(1)Whenpriorknowledgeisnotsufficien,tthere-sultbymeansofsimplerandomsamplingismorereasonable.Butwhenthereissufficientpriorknowledge,stratifiedsamplinginspacecanmakemorerepresentativesamplesforthesampling,andalsocanimprovetheaccuracyofsampling.(2)Thedistributionofsamplepointsofspatialstratifiedsamplingisprimarilyrelatedtostratification.Thehigherthestratifiedleve,l.ie.thehigherthecomplexityoftheregiona,lthemorethenumberofsamplepointsis.(3)Bycomparingtheresultsofspatialstratifiedsamplingbyintegratedfactorsandbyexperts,wecanseethatstratifiedsamplingisdirectlyorindirectlyrelatedwiththestratifyingprocess,therationalityofthestratifyingdesignaffectstherationalityofsamplingresults.So,intheprocessofspatialstratifiedsampling,therationalityofstratifyingdesignisveryimportantforthestratifiedsampling.TheobjectiveofthisstudyistoprovideabasisforthesamplinginthesettlementofXiamenIsland,offeraconceptof/complexity0todescribethesituationofvariousregions,andsuggestseveralideasontheestablishmentofthemult-iobjectivespatialsamplingmode.lKeywords:spatialsampling;mult-iobjectivesampling;socialsamplesurvey;XiamenIsland3642010