基于分层回归遥感抽样技术的冬小麦估产研究_以河北省

      THESISOFMASTERDEGREE                                                                                      09.5.15                                     09.5.15           1MODIS 1. 2. 3. 2 3ABSTRACTTheyielddataisonekeytechnologiesinwinterwheatproductionestimationbyremotesensing,whichplaysafundamentalroleintheentireoperationsofwinterwheatyieldestimationprocess.Gettingthewinterwheatyielddatabeforethecropharvesttimelyandaccuratelyisofgreatsignificancetothenationalfoodsecurityandfoodpolicy-making.Inthispaper,anexampleinShijiazhuangandXingtaiarea,southernHebeiprovince,isgiven,MODISdataisusedtoextractthecoverageareaofwinterwheat,thenremotesensingfactorwhichcouldreflectwinterwheatproductionisselectedasstratifiedbasis,afterautocorrelationtestofwinterwheatsamplepoints,spatialsamplingframeofwinterwheatyieldisconstructed,studyingwinterwheatwithstratifiedregressionsamplingmethod,toinferthemeanyieldofwinterwheatintheoverallstudyareafinally.Themainthesisandconclusionsofthestudyareasfollows:1.Thespatialdistributionofwinterwheatisgotfromtheremotesensingdataofstudyareainthispaper,onthisbasis,samplingunitisbuiltwithremotesensingandgeographicinformationsystemstechnology,afterchoosingstratificationlevelandtestingautocorrelationofsamplepoints,spatialsamplingframeisconstructed,thenthewholesamplingprocessofwinterwheatwithremotesensinginastratifiedregressionsamplingmethodiscompleted.Thestudyshows,thestratifiedregressionsamplingmethodwithremotesensingcanmeettheneedsofstatisticseitherinsamplingaccuracy,samplingefficiencyoreconomicefficiency,whichhasextensiveapplicationprospects.2.Inthispaper,combinedwithremotesensing,geographicinformationsystemsandstatisticalsamplingtechniques,spatialsamplingframeisconstructedonthebasisofremotesensingimages.Testresultsprove,whentheaccuracycouldmeettheneedsofstatistics,spatial-basedsamplingunitandframehasgood4representation,solesssamplesizeisrequiredinthesamplingprocessandthecostofyieldestimationisreduced.3.Autocorrelationisanalyzedamongspatialsamplingunit,sincespatialautocorrelationisanimportantqualityofspatialgeographicdata.Inthispaper,samplingunitisconstructedwithspatialtechnology,afteranalyzingspatialautocorrelationamongspatialsamplingunit,spatialsamplingframeisconstructed,andachievegoodresults.Infuturework,spatialautocorrelationofsamplesneedstobefurtherstudiedtoenhancetherepresentativenessofthesamplepointandoptimizetherationaldistributionofquadrates.Overall,winterwheatyieldestimationbyremotesensingisacomplexprocess.Withsufficientexperimentsthepaperprovesthat,usingthestratifiedregressionsamplingmethodwithremotesensinginwinterwheatyieldestimationcanmeettheneedsofstatistics.Atthesametime,becauseoftimeandexperience,therearestillsomedeficiencieswhichneedtobemodifiedandimprovedinfurtherresearch.KeyWords:stratifiedregressionsamplingmethodwithremotesensingspatialsamplingframespatialautocorrelationyieldestimation567 195%  (LACIE)1984 211()NOAA1998(2004)     3  1.  2. 3.  4MODIS—NDVINDVI1. 2. 3.  1-1  5 1-1(1).MODIS (2). (3). 750m×750m(3×3) (4). (5). (6).  MODIS      6(7). (8). (9). (10). () 20201922Lee205020602070(Landsat)(NPP)208019801986 7(LACIE)110020901991(USGS);1992(MARS)NPP1995(IGBP) 1km;; 150(1100)96111001115()110150() 8GIS140012515019861990125120015GIS208019801983MSS15150209019921993-19962000GIS1:101997-1998 9(2001)(2004)(1994)(2005)(2006)()70(USDA)(NOAA)(NASA) (LACIE) 90%(ChhikaraRS,1986;MacDonaldRB,1980;R.B.MacDonald,1976)1980‐1986,(ARISARS)(AreaSamplingFrame)()NASS(ARMS)NASSNASSARMSNASS1km2(CommitteeonNationalStatistics2007)CharlesPerry(1979)LACIEJiyulChang(2007) 10USDA-NASSMODIS 1996FAO(12)MARS(MonitoringAgriculturebyRemoteSensingTechniques)(60)17MARS(Richards2000)2001LUCAS(LandUse/CoverAreaframeSurvey)(PSUs)1km4,000,0002km1,000,0002003LUCAS(Delincé,2001Bettioetal.2002PascalJACQUES2006)F.J.Gallego(1999)MARSAGRIT(ConsorvioITA,2003Martino,2003)(JapanScienceandTechnologyCorporation)PEPPERS(Projectforestablishmentofplantproductionestimationusingremotesensing)GPS(YoshiakiHONDA,1998;,2006)T.ATsiligrides(1998)HellasSushilPadhan(2001)GISJacquesDeline(2003)LUCASWangDi(2008)JuanJuanJing(2003)NDVIWangJuanle(2008) 11DougRickman(2003)ShunlinLiang(2004)MODISLAIEVIDSSAT8011(,1996;,1998a;,1998b)(MARS)1993 (2000)(2000)·GVG(2008)5500m500m(2007)(2004)LiuHonghui(1999)TMAVHRR(2006)7 128200420057ASTERNDVI83.7%(2004)(2008)WebB/S(2005)GVG  (2002) Ignacio(1974)DongMeiChen(2008)DennisD.Cox(1995) 13(2000)sandwich(2007)IISS(2007)MoranGGMoran(2007)Moran’I(2004)Sandwich(2001)  1.  142. 3. 4.  15114°7′3″115°5′15″36°4′44″37°3′23″2-12-1524304800•d 16600mm5200MJ/m2200d5507-92300-28001310-6728106125  ArcGIS9.2Krasovsky_1940_Albers MODISLansat/MSSLandsat/TM(ETM)SPOT/HRV MODIS(Moderate-resolutionImagingSpectrometer)(EOS)Terra(AM-1)Aqua(PM-1)MODIS2330km1-2200254AquaMODISTerraAqua 17EOSMODIS15MODISMODIS490405-14385nm361314MODIS5%1km≥5006.1MODISAVHRRTM/ETM(1)AVHRRMODIS(2)4(3)36(4)MODIS--HDF(hierarchicaldataformat)12bit0.03NOAA/AVHRR10bit4()(0.60--0.70µrn)()(0.7--1.1µrn)Landsat-TM3(0.63--0.69µm)4(0.76--0.90µm)SPOT-HRVXS2(0.61--0.