简介:Theprocessingmethodappliedtothesidesurfaceisdifferentfromthemethodappliedtothelightpasssurfaceinneodymiumphosphateglass(Nd:glass),andthussubsurfacedefectsremainafterprocessing.ThesubsurfacedefectsinthesidesurfaceinfluencethegainuniformityofNd:glass,whichisakeyfactortoevaluatetheperformanceofamplifiers.Thescatteringcharacteristicsofsidesubsurfacedefectsweresimulatedbyfinitedifferencetimedomain(FDTD)Solutionssoftware.Thescatteringpowersoftheglassfabricatedbyacomputernumericalcontrol(CNC)machinewithoutcladdingweretestedatdifferentincidentangles.Thetrendofthecurvewassimilartothesimulatedresult,whilethesmallestpointwasdifferentwiththecomplextruemorphology.Thesimulationshowedthattheequivalentresidualreflectivityofthecladdingglasscanbemorethan0.1%whenthenumberofdefectsinasinglegriddingisgreaterthan50.
简介:
简介:<正>Someproblemsinusingv-supportvectormachine(v-SVM)forthepredictionofnonlineartimeseriesarediscussed.Theproblemsincludeselectionofvariousnetparameters,whichaffecttheperformanceofprediction,mixtureofkernels,anddecompositioncooperationlinearprogrammingv-SVMregression,whichresultinimprovementsofthealgorithm.ComputersimulationsinthepredictionofnonlineartimeseriesproducedbyMackey-GlassequationandLorenzequationprovidesomeimprovedresults.
简介:students'mathematicalunderstanding,theyarenotableyettomastertothenextdegree.Thedecreasingofstudents'MathqualitycanbeseenonthesurveydataonProgramforInternationalStudentAssessment(PISA)thatconductedeachofthreeyears.On2003,Indonesiawasinthe38^thfrom40stateswiththescore360.On2006,itwasin50thfrom57stateswiththescore391.
简介:Thispaperproposesalearning-basedmethodfortextdetectionandtextsegmentationinnaturalsceneimages.First,theinputimageisdecomposedintomultipleconnected-components(CCs)byNiblackclusteringalgorithm.ThenalltheCCsincludingtextCCsandnon-textCCsareverifiedontheirtextfeaturesbya2-stageclassificationmodule,wheremostnon-textCCsarediscardedbyanattentionalcascadeclassifierandremainingCCsarefurtherverifiedbyanSVM.AlltheacceptedCCsareoutputtoresultintextonlybinaryimage.Experimentswithmanyimagesindifferentscenesshowedsatisfactoryperformanceofourproposedmethod.
简介:TheZ-backlighterlaserfacilityprimarilyconsistsoftwohighenergy,high-powerlasersystems.Z-Beamletlaser(ZBL)(Ramboetal.,Appl.Opt.44,2421(2005))isamulti-kJ-class,nanosecondlaseroperatingat1054nmwhichisfrequencydoubledto527nminordertoprovidex-raybacklightingofhighenergydensityeventsontheZ-machine.Z-Petawatt(ZPW)(Schwarzetal.,J.Phys.:Conf.Ser.112,032020(2008))isapetawatt-classsystemoperatingat1054nmdeliveringupto500Jin500fsforbacklightingandvariousshort-pulselaserexperiments(seealsoFigure10forafacilityoverview).Withthedevelopmentofthemagnetizedlinerinertialfusion(MagLIF)conceptontheZ-machine,theprimarybacklightingmissionsofZBLandZPWhavebeenadjustedaccordingly.Asaresult,wehavefocusedourrecenteffortsonincreasingtheoutputenergyofZBLfrom2to4kJat527nmbymodifyingthefiberfrontendtonowincludeextrabandwidth(forstimulatedBrillouinscatteringsuppression).TheMagLIFconceptrequiresawell-defined/behavedbeamforinteractionwiththepressurizedfuel.HencewehavemadegreateffortstoimplementanadaptiveopticssystemonZBLandhaveexploredtheuseofphaseplates.WearealsoexploringconceptstouseZPWasabacklighterforZBLdrivenMagLIFexperiments.Alternatively,ZPWcouldbeusedasanadditionalfusionfuelpre-heaterorasatemporallyflexiblehighenergypre-pulse.AlloftheseconceptsrequiretheabilitytooperatetheZPWinananosecondlong-pulsemode,inwhichthebeamcanco-propagatewithZBL.Someoftheproposedmodificationsarecompleteandmostofthemarewellontheirway.