简介:MostexistingapplicationsofcentroidalVoronoitessellations(CVTs)lackconsiderationofthelengthoftheclusterboundaries.Inthispaperweproposeanewmodelandalgorithmstoproducesegmentationswhichwouldminimizethetotalenergy—asumoftheclassicCVTenergyandtheweightedlengthofclusterboundaries.TodistinguishitwiththeclassicCVTs,wecallitanEdge-WeightedCVT(EWCVT).TheconceptofEWCVTisexpectedtobuildamathematicalbaseforallCVTrelateddataclassificationswithrequirementofsmoothnessoftheclusterboundaries.TheEWCVTmethodiseasyinimplementation,fastincomputation,andnaturalforanynumberofclusters.
简介:CentroidalVoronoitessellations(CVTs)havebecomeausefultoolinmanyapplicationsrangingfromgeometricmodeling,imageanddataanalysis,andnumericalpartialdifferentialequations,toproblemsinphysics,astrophysics,chemistry,andbiology.Inthispaper,webrieflyreviewtheCVTconceptandafewofitsgeneralizationsandwell-knownproperties.WethenpresentanoverviewofrecentadvancesinbothmathematicalandcomputationalstudiesandinpracticalapplicationsofCVTs.Wheneverpossible,wepointoutsomeoutstandingissuesthatstillneedinvestigating.
简介:Wepresentanovelalgorithmforadaptivetriangularmeshcoarsening.Thealgorithmhastwostages.First,theinputtriangularmeshisrefinedbyiterativelyapplyingtheadaptivesubdivisionoperatorthatperformsaso-calledred-greensplit.Second,therefinedmeshissimplifiedbyaclusteringalgorithmbasedoncentroidalVoronoitessellations(CVTs).TheaccuracyandgoodqualityoftheoutputtriangularmeshareachievedbycombiningadaptivesubdivisionandtheCVTstechnique.Testresultsshowedthemeshcoarseningschemetoberobustandeffective.Examplesareshownthatvalidatethemethod.
简介:AnovelconstructionalgorithmispresentedtogenerateaconformingVoronoimeshforanyplanarstraightlinegraph(PSLG).Itisalsoextendedtotesselatemultiple-intersectedPSLGs.Allthealgorithmsareguaranteedtoconverge.Examplesaregiventoillustrateitsefficiency.
简介:Wetackletheproblemofconstructing2DcentroidalVoronoitessellationswithconstraintsthroughanefficientandrobustconstructionofboundedVoronoidiagrams,thepseudo-dualoftheconstrainedDelaunaytriangulation.WeexploitthefactthatthecellsoftheboundedVoronoidiagramcanbeobtainedbyclippingtheordinaryonesagainsttheconstrainedDelaunayedges.Theclippingitselfisefficientlycomputedbyidentifyingforeachconstrainededgethe(connected)setoftriangleswhosedualVoronoiverticesarehiddenbytheconstraint.TheresultingconstructionisamenabletoLloydrelaxationsoastoobtainacentroidaltessellationwithconstraints.
简介:AssumingthataVoronoidiagramofsliceareaisobtained,topologicalstructuresofallVoronoiedgesandVoronoipolygonsareusedtoacceleratetheoffsettingprocess.OncewalklineintersectswithoneofVoronoiedgesofthestartingVoronoiobject,thenextstartingVoronoiobjectisacquiredthroughthetopologyrelationship.Experimentalresultsshowtheapproachiseffectiveandsimple.
简介:Therearemanyphenomenathatgeneratepolygonaltessellationsonsurfacesof3Dobjects.Oneinterestingexampleisthejackfruit,amultiplefruitfoundinthetropics.Arecentstudyfoundthebest-fitsphericalVoronoidiagramfromaphotoofjackfruitskin,buttheoptimizationwasrelativetotheradiusofthesphereandtheheightofthespikes.Inthisstudy,weproposeamethodforadjustingthepositionofthecenterofthesphereinadditiontotheseparameters.Experimentswereconductedusingbothidealandrealdata.However,convergencewithrealdatahasnotbeenconfirmedduetorelaxationoftheconvergencecondition.
简介:Acontourtreeisagoodgraphicaltoolforrepresentingthespatialrelationsofcontourlinesandhasmanyapplicationsinmapgeneralization,mapannotation,terrainanalysis,etc.AnewmethodforgeneratingcontourtreesbyintroducingaVoronoi-basedinterioradjacencyconceptisproposedinthispaper.Theimmediateinterioradjacencysetisemployedtoidentifyallofthechildrencontoursofeachcontourwithoutcontourelevations.Ithasadvantagesoverexistingmethodssuchasthegeometricmethodandtheregiongrowingbasedmethod.
简介:[1]AbdelmotyAI,WilliamsMH.Advancedgeographicdatamodeling.Spatialdatamodelingandquerylanguagefor2Dand3Dapplications.Delft:TheNetherlands,1994[2]FranzA.Voronoidiagrams——asurveyofafundamentalgeometricdatastructure.ACMComputingSurveying,1991,23(3)[3]ZhengC,ChenJ.Organizingandretrievingcadasatraldatabasedonspatio-temporaltopologybetweenfather-sonparcels.JournalofWuhanTechnicalUniversityofSurveyingandMapping,1997,22(3):216~221(inChinese)[4]ChenJ,GoldCM,CuiBL,YongHU,etal.Extendingordinaryplanar-graph-basedspatialdatamodelwithVoronoiapproach.In:ProceedingsofIEAS&IWGIS'97.Beijing,1997.18~20[5]ChenJ,CuiBL.AddingtopologicalfunctionstoMapInfowithVoronoiapproach.JournalofWuhanTechnicalUniversityofSurveyingandMapping,1997,22(3):195~200(inChinese)[6]Clementini,Elisel,PaolinoDiFelice,PetervanOosterom.Asmallsetofformaltopologicalrelationshipssuitableforend-userinteraction.AdvancesinSpatialDatabases,LectureNotesinComputerScience.Springer-Verlag,1993.277~295[7]EdwardsG,LigozatG,FrylA,etal.AVoronoi-basedPIVOTrepresentationofspatialconceptsanditsapplicationtoroutedescriptionsexpressedinnaturallanguage.SpatialDataHandling'96,1996,7B.1~15[8]EgenhoferMJ.Reasoningaboutbinarytopologicalrelations.In:Proceedingsofthe2ndSymposiumonLargeSpatialDatabases.LectureNotesinComputerScience,Springer-Verlag,1991.143~160[9]EgenhoferMJ,FranzosaRD.Point-settopologicalspatialrelations.INT.J.GeographicalInformationSystems,1991,5(2):161~176[10]EgenhoferMJ,HerringJ.Categorizingbinarytopologicalrelationshipsbetweenregions,lines,andpointsingeographicdatabases.TechnicalReport,DepartmentofSurveyingEngineering,UniversityofMaine,Orono,1991[11]EgenhoferMJ,Al-tahaKK.Reasoningaboutgradualchangesoftopologicalrelations,inTheoriesandMet
简介:Thispaperconsidershowtouseagroupofrobotstosenseandcontroladiffusionprocess.Thediffusionprocessismodeledbyapartialdifferentialequation(PDE),whichisabothspatiallyandtemporallyvariantsystem.Therobotscanserveasmobilesensors,actuators,orboth.CentroidalVoronoiTessellationsbasedcoveragecontrolalgorithmisproposedforthecooperativesensingtask.Forthediffusioncontrolproblem,thispaperconsiderssprayingcontrolviaagroupofnetworkedmobilerobotsequippedwithchemicalneutralizers,knownassmartmobilesprayersoractuators,inadomainofinteresthavingstaticmeshsensornetworkforconcentrationsensing.ThispaperalsointroducestheinformationsharingandconsensusstrategywhenusingcentroidalVoronoitessellationsalgorithmtocontroladiffusionprocess.Theinformationissharednotonlyonwheretospraybutalsoonhowmuchtosprayamongthemobileactuators.BenefitsfromusingCVTandinformationconsensusseekingforsensingandcontrolofadiffusionprocessaredemonstratedinsimulationresults.
简介:InordertoquantitativelyevaluatethespuriousdianeutralmixinginaglobaloceanmodelMPAS-Ocean(ModelforPredictionAcrossScales)usingasphericalcentroidalvoronoitessellationsdevelopedjointlybytheNationalCenterforAtmosphericResearchandtheLosAlamosNationalLaboratoryintheUnitedStates,wechoosez*verticalcoordinatesysteminMPAS-Ocean,inwhichallphysicalmixingprocesses,suchasconvectionadjustmentandexplicitdiffusionparameterschemes,areomitted,usingalinearequationofstate.BycalculatingtheReferencePotentialEnergy(RPE),frontrevolutionposition,timerateofRPEchange,probabilitydensityfunctiondistributionanddimensionlessparameterχ,fromtheperspectivesofresolution,viscosity,HorizontalGridReynoldsNumber(HGRN),Re?,andmomentumtransmissionscheme,usingtwoidealcases,overflowandbarocliniceddychannel,wequalitativelyanalyzethesimulationresultsbycomparisonwiththethreenon-isopycnalmodelsinIlicaketal.(2012),i.e.,MITGCM,MOM,andROMS.TheresultsshowthatthespuriousdianeutralmixingintheMPAS-Oceanincreasesovertime.ThespuriousdianeutraltransportisproportionaltotheHGRNdirectlyandisreducedbyincreasingthelateralviscosityorusingafinerresolutiontocontrolHGRN.WhentheHGRNislessthan10,spurioustransportisreducedsignificantly.Whenusingtheproperviscosityclosure,MPAS-OceanperformsbetterthanMITGCMandMOM,closelytoROMS,inthe2Dcasewithoutrotation,andmuchbetterthantheabove-mentionedthreeoceanmodelsundertheconditionof3Dspacewithrotationduetothecellareadifferencebetweenthehexagoncellandthequadrilateralcellwiththesameresolution.BoththeZalesak(1979)fluxcorrectedtransportschemeandLeithclosureinMPAS-Oceanplayanexcellentroleinreducingspuriousdianeutralmixing.TheperformanceofLeithschemeispreferabletotheconditionofthree-dimensionalbarocliniceddy.
简介:ThispaperpresentsthenumericalinvestigationofthedistributionofinertialsedimentparticlesintheturbulentboundarylayerofanopenchannelflowwiththeparticleStokesnumberrangingfrom0.6to20.4.Themethodologyisacombinationofthreenumericalapproaches,i.e.directnumericalsimulationofturbulentflow,thepoint-particleimmersedboundarymethod,andthediscreteparticlemethod.ByapplyingtheVorono?analysis,thepreferentialconcentrationcharacteristicsofsedimentparticleswereinvestigatedquantitatively.ItwasfoundthatthenormalizedareaoftheVorono?cellsfollowsalognormalparticledistribution.Theinertialsedimentparticlesdistributedunevenlyintheturbulentboundarylayerandtheunevenness,governedbytheparticleStokesnumber,wasmoresignificantastheparticleStokesnumberapproachesunity.Theinertialsedimentparticlesintheturbulentboundarylayeraccumulatedpreferentiallyinstreamwise-alignedstreakystructuresandthispatternwaslesssignificantwithincreasingparticleStokesnumber.