基于高分辨率遙感影像中艦艇目標的分割與提取技術研究.pdf

1、東北師范大學碩士學位論文基于高分辨率遙感影像中艦艇目標的分割與提取技術研究姓名：李明申請學位級別：碩士專業(yè)：地圖學與地理信息系統(tǒng)指導教師：楊令賓20080501Abstractln陽motesensingimage，these呻entatiOnofmmtarytar鼯ts，featuleext均ctionandthetechnolol叮ofautomat：icIeco衛(wèi)nitionaIethehotspotanddif五cul鉀inth

2、e心searchofmode髓mmtaryinfo珊ationt鉀hnoIo黟Theya聆appliedmainlyinmany6eldssuchasthecollectionofmilitaryinfbmation，monitoringandwarningagainsttheenemiesandhomingl舛idanceweaponsInhighr霉solution《WhichinvolvespatiaI代solution，temp

3、o豫l心soIutionandspectmm№soIution)satelIiteimage，theimagepmcessingand心cognitiontechnolo留ofshiptarg橢haveag№atVaIueinmemnitaryp心pamtionagainstThiwanandsove托i驢鑼pmblemsinEastSeaandSouthChinaSeaTheyprovideuswithasi驢incantinfo珊a

4、tionsafe霉mardnetonOuroceanicsovereign鑼，nauticaIinteI℃stsandnauticaIbatltIesagainstenemiesMeanwhile，inthemodemizati帆ofourmilitary，theyanurgen呵withthetransitiOnfommechanizationtoinflormationizationThesatellitenmotesensingi

5、magesfhmthepapera他thevismleima譬etakenbyhomemade“JBⅢ’’militaryncon的issancesatelliteThespatiaI陽solutionamng懿fbmlto3metersThispaper∞ncent豫teson伽ep№cisionse2mentationinshipta曜etsofthe他motesensingima酗andthefeatIl他ext豫ctionwhi

6、chcomesfhm伯eshiptargetsafterse腫蛐tationCombiningwiththefactIIaIcha隨cteristicsof協(xié)eshiptargetsandbacl‘gmundofthewatebthepapern礴tlycarrythmughayawp№moValandpIetreatmentofcontrastenhancement，foIIowedbyanimpIo、rementonthecurre

7、ntfhzzyenhancementalgorithminthepIocessingofimageenhancementInthefoHowingpmcessOfimagesegmentation，thepaperfo伽s鶴onimagesegmentationmethodwhichisbasedonthIeshold，thenmakesansegmentationexperimenta2ainstshiptargetsbymeanso

8、fwen—appHedonedimensionalthlesholdsegmentationmethod，afterthatanalysesthedi婦dvantageItspreadsthefeatu心spaceofthedistributionpixelsfIomonetotwodimensionbymakinguseofgray—levelinfo冊ationandspatialinfbrmationintheprocessOfs

9、egmentatiOnIt主ndicat銘a璺DOdr鶴ultbyadoptingthetraditionaltwodimentionaIotsuthIesholdsegmentationtosegmentthetargetswhiletheshortcomingisthatithasacompl旺ingcomputationandistimeconsumingFinalb’th蛔paperp№sentsamethodthatcombi

10、nedwith代giongmwingmethod，itsegmentsimagesbymeansofal‘indofsimpIifiedtwo—dimensionalotsuThemethoddividestheima2esegmentationinto鉚op1ocessesFirstly，ittakesainitiaIsegmentationbyusingsimpIi矗edtwO—dimentiOnalotsuSecondly，itc

11、alriesonaprecisiOnsegmentatiOnbyusingregiOn—gn時yingmethodExperimentalresuItindicatesthatthemethodpIovestobeaccuraterapidandwell—appliedThepaperismadeupofflvechapte礴，thekeypointofwhicharetwo，th聆eandfourchapteLTheseth心echa

12、pte礴make№search心spect沁elyonthetechnolo留ofimageprepmcessing，imagesegmentationandfeatIlreextraction，impmVementandexperimentwhichnfb聆dinthispaperKeywords：high托solution、imagesegmentation、featlInextraction、th№shoId、他昏ong加wing