基于机载高光谱遥感数据的溢油信息提取方法-刘丙新-张志达-李颖-陈澎

1[e-mail]．401Vol．40No．120142JournalofDalianMaritimeUniversityFeb．20141006-7736201401-0089-04116026．MNFMNF．．TP751．1AExtractionmethodofoilspillinformationusingairbornehyper-spectralremotesensingdataLIUBing-xinZHANGZhi-daLIYingCHENPengNavigationCollegeDalianMaritimeUniversityDalian116026ChinaAbstractDecisiontreeclassificationmethodwasproposedonbasisoftheminimumnoisefractionMNFtoreducedimen-sionsofhyper-spectralremotesensingdataandimprovepro-cessingefficiency．ThedataredundancywasreducedbymeansofMNFandthefigurenoisewasseparated．Thedeci-siontreewasestablishedaccordingtoanalyzinglandmarks’MNFeigenvalueandtherelativethicknessoftheoilfilmwasextracted．Theresultsshowthatthemethodmentionedcouldensurerecognitionaccuracyachieveeffectiveuseofspectraldimensioninformationaswellasreducetheprocessingtimesignificantlysoastomakethequickproductsforoilspillre-sponsebyusinghyper-spectraldata．Keywordshyper-spectralremotesensingoilspillmonito-ringminimumnoisefractionMNF01．、2．．、3．SAM4、56．．MNFMNFMNF．1GulfofMexico．2010425“”DWH120107151970～43009040“”7－8．/AVIＲIS380～2500nm22410nm9．201079AVIＲIS10TwinOtterN70AＲDWH5km—3．2m．1“”Fig．1Positionofdeepwaterhorizonplatform22．1MNFMNF2．、、MNFMNFNMNFMNFMNF．2．2Green11．MNF12．2MNFFig．2FlowchartofclassificationusingdecisiontreebasedonMNF．CN=∑NDN=UTCNU1DNCNU．1P=UD－1/2NI=PTCNP2I．PXY=PX．．PCDCD－objPTCDP=CD－obj3CD－objVVTCD－objV=DD－objDD－objVVTV=II．MNFTMNF=PVMNF．19133．1AVIＲIS3．、、、sheen395～531nm．531nm．521～19nm．、．3Fig．3Typicalobjectsandtheextractedspectra3．2MNFMNF254MNF．MNF．MNF、、、MNF、、MNF0MNF0、Band2＞Band1MNFBand1＞Band2MNF0MNF0MNF．4MNFFig．4MNFeigenvalueoftypicalobjects4MNF5．5MNFFig．5ClassificationdecisiontreebasedonMNFMNF669．16%12．26%．5．50%6．22%．DWH、．5MNF92406Fig．6ＲesultofclassificationpositionofDeepWaterHorizonplatform．MNF．．Ｒeferences1SOLBEＲGAHS．Ｒemotesensingofoceanoil－spillpollutionJ．ProceedingsoftheIEEE201210010SI2931－2945．2BＲEKKECSOLBEＲGAHS．OilspilldetectionbysatelliteremotesensingJ．ＲemoteSensingOfEnviron-ment2005951－13．3．———、M．2006．4SALEMFMF．Hyperspectralremotesensinganewap-proachforoilspilldetectionandanalysisD．VirginiaUnitedStatesGeorgeMasonUniversity2003．5SANCHEZGＲOPEＲWEGOMEZＲ．DetectionandmonitoringofoilspillsusinghyperspectralimageryJ．Geo-SpatialandTemporalImagesandDataExploitationIII20035097233－240．6PLAZAJPＲEZＲPLAZAAetal．Mappingoilspillsonseawaterusingspectralmixtureanalysisofhy-perspectralimagedataJ．ChemicalandBiologicalStandoffDetectionIII2005599591－98．7JOYESBMACDONALDIＲLEIFEＲIetal．Magni-tudeandoxidationpotentialofhydrocarbongasesreleasedfromtheBPoilwellblowoutJ．NatureGeoscience201143160－164．8SVEJKOVSKYJLEHＲWMUSKATJetal．Opera-tionalutilizationofaerialmultispectralremotesensingduringoilspillresponselessonslearnedduringthedeepwaterhorizonMC－252spillJ．Photogrammet-ricEngineeringandＲemoteSensing201278101089－1102．9GＲEENＲOEASTWOODMLSAＲTUＲECM．Ima-gingspectroscopyandtheairbornevisible/infraredima-gingspectrometerAVIＲISJ．ＲemoteSensingofEn-vironment199865227－248．10BＲADLEYESＲOBEＲTSDADENNISONPE．GoogleearthandGooglefusiontablesinsupportoftime-criticalcollaborationmappingthedeepwaterhorizonoilspillwiththeAVIＲISairbornespectrometerJ．EarthScienceInformatics20114169－179．11GＲEENAABEＲMANMSWITZEＲPetal．AtransformationfororderingmultispectraldataintermsofimagequalitywithimplicationsfornoiseremovalJ．IEEETransactionsonGeoscienceandＲemoteSensing199826165－74．12．MNFSVMJ．2007512－15．LIHai-taoGUHai-yanZHANGBingetal．ＲesearchonhyperspectralremotesensingimageclassificationbasedonMNFandSVMJ．ＲemoteSensingInforma-tion2007512－15．inChinese