摘要
Shorelineextractionisfundamentalandinevitableforseveralstudies.Ascertainingtheprecisespatiallocationoftheshorelineiscrucial.Recently,theneedforusingremotesensingdatatoaccomplishthecomplextaskofautomaticextractionoffeatures,suchasshoreline,hasconsiderablyincreased.Automatedfeatureextractioncandrasticallyminimizethetimeandcostofdataacquisitionanddatabaseupdating.Effectiveandfastapproachesareessentialtomonitorcoastlineretreatandupdateshorelinemaps.Here,wepresentaflexiblemathematicalmorphology-drivenapproachforshorelineextractionalgorithmfromsatelliteimageries.Thesalientfeaturesofthisworkarethepreservationofactualsizeandshapeoftheshorelines,run-timestructuringelementdefinition,semi-automation,fasterprocessing,andsinglebandadaptability.Theproposedapproachistestedwithvarioussensor-drivenimageswithlowtohighresolutions.Accuracyofthedevelopedmethodologyhasbeenassessedwithmanuallypreparedgroundtruthsofthestudyareaandcomparedwithanexistingshorelineclassificationapproach.Theproposedapproachisfoundsuccessfulinshorelineextractionfromthewidevarietyofsatelliteimagesbasedontheresultsdrawnfromvisualandquantitativeassessments.
出版日期
2017年04月14日(中国期刊网平台首次上网日期,不代表论文的发表时间)