学科分类
/ 2
31 个结果
  • 简介:Thispaperpresentsanovelapproachtofeaturesubsetselectionusinggeneticalgorithms.Thisapproachhastheabilitytoaccommodatemultiplecriteriasuchastheaccuracyandcostofclassificationintotheprocessoffeatureselectionandfindstheeffectivefeaturesubsetfortextureclassification.Onthebasisoftheeffectivefeaturesubsetselected,amethodisdescribedtoextracttheobjectswhicharehigherthantheirsurroundings,suchastreesorforest,inthecoloraerialimages.Themethodologypresentedinthispaperisillustratedbyitsapplicationtotheproblemoftreesextractionfromaerialimages.

  • 标签: 遗传算法 特征提取 分类结构 图像提取 交叉 变化
  • 简介:Stabilityparameters(Monin-ObukhovlengthL,gradientRichardsonnumberRiandbulkRischardsonnumberRi),whichareapplicableinurbanenvironment,werediscussedforwaysofcalculatingclassificationstandards.Gradientobservationsfroma325-mmeteorologicaltowerinBeijingareusedtocategorizeRibbasedonthreedifferentstandardsofstabilityproposedbyD.Golder,IrwinandHoughton.TheresultsshowthatitisrelativelyreasonablefortheregionofBeijingtoapplytheclassificationstandardbyIrwin.

  • 标签: 北京 大气探测 数据处理 稳定性
  • 简介:ObjectiveDebrisflowsarecohesivesedimentgravityflowswhichoccurinbothsubaerialandsubaqueoussettings.Comparedtosubaerialdebrisflowswhichhavebeenwellstudiedasageologicalhazard,subaqueousdebrisflowsshowingcomplicatedsedimentcompositionandsedimentaryprocesseswerepoorlyunderstood.Themainobjectiveof

  • 标签:
  • 简介:Anobjectiveanalysisoftropicalcyclonetracksisperformed,withwhichthetracksof131tropicalstorms(TSs)in1972-2011areseparatedintothreetypesthatmovewest-,north-andnorthwestward,denotedasTypesA,BandC,respectively.TypeA(21TSsand16%oftotal)hastheorigininthesouthwesternBayofBengal,withtheTSinaunimodaldistributionasitsseasonalfeature,occurringmainlyinautumn;18ofthe21TSs(takingup90%)landmostlyonthewesternBaycoast(westof85°E);5%ofType-ATSsattainsthewindspeedof>42.7to48.9m/s.TypeAhaslittleornoeffectonTibet.TypeB(74TSs,56.6%ofthetotal)hasitspreferableorigininthecentralBayofBengal,withtheTSinabimodaldistributionasitsseasonalpattern.Thistypedenotesthetravelinthenorthinspring,withthelandfallof67ofthe74TSs(accountingfor91%)mainlyonthemiddlecoastoftheBay(85°to95°E),and19%oftheTSsreachingthewindvelocityof>42.7to48.9m/s,whichexertgreateffectonTibetanditisthisTStrackthatgivesstrongprecipitationonitswaythroughthisregion.TypeC(36TSs,27.5%ofthetotal)hasitsmainorigininthesouthernpartofthebay,andtheseTSsareformedlargelyinautumn,movinginthenorthwestdirection,and23ofthe36TSs(64%)landmostlyonthewesternBaycoast,lastingforalongertime,withalmostnoimpactuponTibet.

  • 标签:
  • 简介:Themultibeamsonarscanprovidehydrographicqualitydepthdataaswellasholdthepotentialtoprovidecalibratedmeasurementsoftheseaflooracousticbackscatteringstrength.Therehasbeenmuchinterestinutilizingbackscattersandimagesfrommultibeamsonarforseabedtypeidentificationandmostresultsareobtained.Thispaperhaspresentedafocusedreviewofseveralmainmethodsandrecentdevelopmentsofseafloorclassificationutilizingmultibeamsonardataor/andimages.Theseareincludingthepowerspectralanalysismethods,thetextureanalysis,traditionalBayesianclassificationtheoryandthemostactiveneuralnetworkapproaches.

  • 标签: 海底分类法 多束声纳 反散射度 探测鉴别
  • 简介:在这研究,Jenkinson和Collison(1977)基于兰姆(1950)的一个打字计划开发的分类计划被使用在一个每月的基础上从吝啬的海平面的压力获得发行量类型。从1951~2002的每月吝啬的海平面的压力数据被用来导出六个发行量索引并且向一个发行量目录提供27种发行量类型。最经常发生了的五种主要类型(N,NW,C,CSW,和SW)被分析在各种各样的时间规模上与哈尔滨的温度揭示他们的关系。逐步的多重回归被用来重建温度异例。发生的所有类型和三个学生的合成地图的每月吝啬的降雨打字(C,CSW,和SW)与哈尔滨的降水相关被学习。结果证明在冬季的主导的类型是类型N和NW。类型C,CSW,和SW在夏天经常发生。类型N和NW赞成一个否定温度异例并且当类型C,CSW,和SW经常导致一个积极温度异例并且对应于更多的降雨时,对应于更少的降雨。而且,一个成功的统计模型能被建立与仅仅,六之一索引并且大规模吝啬的温度。用模型,在在1951和2002之间的温度异例的77.3%全部的变化能被重建。类型C与全部的降雨有一种靠近的关系,类型C降水在在最近的年里决定哈尔滨的全部的降雨起一个主要作用。这个分类计划是有温度的一个统计downscaling模型和它的关系,降水能被用来预报地区性的气候。[出版摘要]

  • 标签: 月平均大气环流型 分型方法 温度 降水 哈尔滨 气候
  • 简介:Thispaperpresentsanewwaytoextractconceptthatcanbeusedtoimprovetextclassificationper-formance(precisionandrecall).Thecomputationalmeasurewillbedividedintotwolayers.Thebottomlayercalleddocumentlayerisconcernedwithextractingtheconceptsofparti-culardocumentandtheupperlayercalledcategorylayeriswithfindingthedescriptionandsubjectconceptsofparticularcategory.Therelevantim-plementationalgorithmthatdramatic-allydecreasesthesearchspaceisdis-cussedindetail.Theexperimentbasedonreal-worlddatacollectedfromInfo-Bankshowsthattheapproachissupe-riortothetraditionalones.

  • 标签: 概念 计算方法 运算法则 正文 分类 有效性
  • 简介:一个模糊ARTMAP分类器为分类实验ofCBERS-2形象被采用。基本理论并且关于算法处理首先被介绍,在CBERS-2高决定形象上在Shihezi县与一个陆地使用分类实验列在后面。三个分类器被比较:最大的可能性分类器(MLC),错误背繁殖(BP)分类器,和模糊ARTMAP分类器。比较比MLC和BP的高为模糊ARTMAP分类器显示出可比较地更好的结果,与9.9%和4.6%的全面分类精确性。结果也证明模糊ARTMAP分类器在在CBERS-2形象上识别赤裸的土壤有更好的洞察力。

  • 标签: 模糊ARTMAP分类器 CBERS-2图象 分类实验 土地利用
  • 简介:SequenceboundaryisstudiedindetailinreferencetothePermianGuadalupianLopingianboundary,basedonhighresolutioncorrelationofconodootzones.Aconceptualsynchronousstratigraphicboundary,correspondingtotheageofthesequenceboundarydatedintheconformableportionofthesequenceboundary,isdefinedasthesequencechronostratigraphicboundary(SCB).Thesequenceboundaryisprovedtobedualinnatureinregardstothesynchronyanddiachroneityofstratigraphicboundaries.Themeritsofthesequenceboundaryinstratigraphicsubdivision,correlationandclassificationareelucidatedincomparisonwiththelithostratigraphicboundary,thebiostratigraphicboundaryandthetraditionalapproachoftheglobalstratotypesectionandpoint(GSSP).ByintegratingthevirtuesofthesequenceboundaryandtheGSSP,itisproposedthattheGSSPshouldbeestablishedintheconformableportionoftherelatedsequenceboundary.Theboundaryestablishedinlightofthisapproachisdefinedas

  • 标签: SEQUENCE BOUNDARY SEQUENCE chronostratigraphic BOUNDARY -
  • 简介:基于地面的云分类由于在在不同大气的条件下面的云的外观的极端变化是挑战性的。质地分类技术最近被介绍了处理这个问题。一个新奇质地描述符,突出的本地二进制模式(SLBP),为基于地面的云分类被建议。SLBP利用最经常发生的模式(突出的模式)捕获描述的信息。这个特征使SLBP柔韧到噪音。用基于地面的云图象的试验性的结果证明建议方法能比当前的最先进的方法完成更好的结果。

  • 标签: 云分类 二进制模式 突出 陆基 大气条件 分类技术
  • 简介:BasedontheJointTyphoonWarningCenter(JTWC)best-trackdatasetbetween1965and2009andthecharacteristicparametersincludingtropicalcyclone(TC)position,intensity,pathlengthanddirection,amethodforobjectiveclassificationoftheNorthwesternPacifictropicalcyclonetracksisestablishedbyusingk-meansClustering.TheTClifespan,energy,activeseasonandlandfallprobabilityofsevenclustersoftropicalcyclonetracksarecomparativelyanalyzed.Thecharacteristicsoftheseparametersarequitedifferentamongdifferenttropicalcyclonetrackclusters.Fromthetrendofthepasttwodecades,thefrequencyofthewesternrecurvingcluster(accountingfor21.3%ofthetotal)increased,andthelifespanelongatedslightly,whichdiffersfromtheotherclusters.TheannualvariationofthePowerDissipationIndex(PDI)ofmostclustersmainlydependedontheTCintensityandfrequency.However,theannualvariationofthePDIinthenorthwesternmovingthenrecurvingclusterandthepelagicwest-northwestmovingclustermainlydependedonthefrequency.

  • 标签: tropical cyclone classification of tracks K-means clustering character of cluster
  • 简介:Interestingclassificationsofbasinogenesisandbasinswereproposedbymanyscientists.Theyclassifiedbasinogenesisandbasinsmainlyfromasingleangle,eitherfromahistoricalangleorfromadynamicangle.Inordertomorecomprehensivelyunderstandthemformoreeffectivelyguidingprospectingandexplo-ration,theauthorintegratesthetwomethodsofanalysiswitheachotherandproposesanintegrativeclassification.Accordingtothehistorical-dynamicintegrativeclassification,basinogenesisandbasinscanbe_di-videdintothreetypes:oceaniccrusttype,embryo—continental(transitional)crusttypeandcontinentalcrusttype.Oceaniccrusttypecanbesubdividedintomobileregiontype(mainlytensional)andstableregiontype.Embryo—continentaltypeincludespre-geosynclinaltype(divisibleintoseveralmobileregiontypesandstableregiontypeswithtensionaltypepredominatingamongmobileregiontypes)andear-ly-geosynclinaltype(mainlytensional).Continentalcrusttypeincludeslategeosynclinal(fo

  • 标签: basinogenesis and ORE-FORMING basin historical-dynamic INTEGRATIVE
  • 简介:专家知识广泛地被用来改进遥远地察觉到的分类精确性。通常,专家分类系统主要取决于DEM和一些题目的地图。在象素水平的空间关系信息通常被介绍进专家分类。因为地理目标被发现关系空间地依赖到某个度,在象素的空间关系信息的通常使用的基本单位极大地限制了空间信息的效率。寻找算法的一位基于补丁的邻居被建议实现专家分类。同类的光谱的联合起来,补丁,在空间目标被用作基本单位补丁鈥?关系信息的颗粒度,和不同类型通过寻找的一位空间邻居被获得算法。然后邻居信息和DEM数据被增加进专家分类系统并且过去常修改原始分类错误。在这,情况,沼泽地的分类精确性,草地和农田显然被改进。在这个工作,当水抽取方法的基础目标,和不同类型被测试在高精确性得到结果,水被使用。

  • 标签: 曲面 邻域搜索算法 专家分类法 空间关系 空间飞行器 遥感
  • 简介:Onthebasisofthecarbonisotopiccompositionsofmethane(CH4)anditshomologuesandthedifferencesinisotopicvaluesforCH4andethane(C2H6)andthecorrelationandcompositionalchar-acteristicsofhydrocarbongases,theauthorhasproposedageneticclassificationofnaturalgasesintheoil-gaszone.Theyareclassifiedasbiogeneticandabiogeneticgasesintermsofthetypesofhydrocarbon-generatingprecursors(orparentmaterials)andtheirthermalevolutionstages.Biogeneticgasescanalsobefurtherdividedintotwoseries:biochemicalandthermochemicalgases,withthelat-terformedatdifferentevolutionstages.Gasesgeneratedfromtype-Iand-II1organicmatterarecalledoil-seriesgases,thosefromtype-III,coal-series,andthosetype-II2,mixture-typegases.Gasesgeneratedfromtwoormorethantwotypesofprecursorsarecalledmixture-sourcegases.Accordingtothosementionedabove,naturalgasesfromthemajoroil-gaspoolsintheSichuanBasinhavebeendiscriminantlyanalyzed,andtheresultsareconcordantwiththedistributionandde-velopmentofhydrocarbon-sourcerocksaswellaswiththeircharacteristics,indicatingaprospectiveapplication.

  • 标签: 四川盆地 油气藏 甲烷 烃源岩 碳同位素 有机物
  • 简介:使用是的地震属性展示因为在象地震外形分析那样的各种各样的目的,在特征空间的分类为地震解释的目的是常规的。但是有时地震的数据可以没有属性或它是难的在一些应用程序定义属性的一个小、相关的集合。因此,采用执行外形没有使用属性,当模特儿的技术是必要的。在这份报纸我们在场为外形与错过当模特儿地震数据的一个新方法归因那叫的不同基于的分类。在这个方法,分类基于当模特儿的不同和外形将在不同空间被做。在这个空格,不同考虑同样新的特征而不是真实特征。一个强大的分类器在两个特征空格被采用的一台支持向量机器(基于特征)并且不同空间(无特色)为外形分析。建议无特色、基于特征的分类从一块伊朗的油地在一个真实地震数据上被使用。当模特儿的使用的地震属性更好提供的外形结果,但是无特色的分类结果也是令人满意的,外形关联是可接受的。确实,属性的力量引起区别不同外形到分析使用属性提供更可靠的结果比较给无特色的外形分析的那外形。

  • 标签: 地震属性 地震资料 分类器 相异性 建模 地震相分析
  • 简介:Buildingcodeshavewidelyconsideredtheshearwavevelocitytomakeareliablesubsoilseismicclassification,basedontheknowledgeofthemechanicalpropertiesofmaterialdepositsdowntobedrock.Thisapproachhaslimitationsbecausegeophysicaldataareoftenveryexpensivetoobtain.Recently,otheralternativeshavebeenproposedbasedonmeasurementsofbackgroundnoiseandestimationoftheH/Vamplificationcurve.However,theuseofthistechniqueneedsaregulatoryframeworkbeforeitcanbecomearealisticsiteclassificationprocedure.ThispaperproposesanewformulationforcharacterizingdesignsitesinaccordancewiththeAlgerianseismicbuildingcode(RPA99/ver.2003),throughtransferfunctions,byfollowingastochasticapproachcombinedtoastatisticalstudy.Foreachsoiltype,thedeterministiccalculationoftheaveragetransferfunctionisperformedoverawidesampleof1-Dsoilprofiles,wheretheaverageshearwave(S-W)velocity,Vs,insoillayersissimulatedusingrandomfieldtheory.Averagetransferfunctionsarealsousedtocalculateaveragesitefactorsandnormalizedaccelerationresponsespectratohighlighttheamplificationpotentialofeachsitetype,sincefrequencycontentofthetransferfunctionissignificantlysimilartothatoftheH/Vamplificationcurve.ComparisonisdonewiththeRPA99/ver.2003andEurocode8(EC8)designresponsespectra,respectively.Intheabsenceofgeophysicaldata,theproposedclassificationapproachtogetherwithmicro-tremormeasurescanbeusedtowardabettersoilclassification.

  • 标签: RANDOM field TRANSFER function soil classification
  • 简介:IthaslongbeenacknowledgedthatGISdatacanbeusedasauxiliaryinformationtoimproveremotesensingimageclassification.Inpreviousstudies,GISdatawereoftenusedintrainingareaselectionandpostprocessingofclassificationresultoractedasadditionalbands.Generally,itisfulfilledinastatisticalorinteractivemanner,soitisdifficulttousetheauxiliarydataautomaticallyandintelligently.  Furthermore,iftheclassifierrequestscertainstatisticalcharacteristics,theadditionalbandmethodcannotbeusedbecausemostauxiliarydatadonotmeettherequirementsofstatisticalcharacteristics.Ontheotherhand,expertsystemtechniqueswereincorporatedinremotesensingimageclassificationtomakeuseofdomainknowledgeandlogicalreasoning.Butbuildinganimageclassificationexpertsystemwasverydifficultbecauseofthe“knowledgeacquisitionbottleneck”.  Spatialdataminingandknowledgediscovery(SDMKD),istheextractionofimplicit,interestingspatialornon_spatialpatternsandgeneralcharacteristics.Weproposedatheoreticalandtechnicalframeworkofspatialdataminingandknowledgediscovery(Lietal.,1997).Andspatialdataminingissupposedtobeusedintwoaspects,oneisintelligentanalysisofGISdata,theotheristosupportknowledgedriveninterpretationandanalysisofremotesensingimages.SDMKDprovidesanewwayofknowledgeacquisitionforremotesensingimageclassification.Severalresearchershavedonesomeworkinthisfield.Eklundetal.(1998)extractedknowledgefromTMimagesandgeographicdatainsoilsalinityanalysisusinginductivelearningalgorithmC4.5.Huangetal.(1997)extractedknowledgefromGISdataandSPOTmultispectralimageinwetlandclassificationusingC4.5too.Inthesetwostudies,geographicdatawereconvertedfromvectortorasterformatinwhichthesamplingsizeisequaltoimagepixelsize.Theimplementationofdataminingtechniquesinspatialdatabase,especiallyinductivelearningmethod,andthecombinationo

  • 标签: data MINING KNOWLEDGE DISCOVERY image classification