基于现代优化算法曲线识别方法-应用数学专业论文.docx

Curveidentificationmethodbasedonmodernoptimization algorithmAbstractCurveidentificationisthebasictaskaboutthefieldinimagematchingandidentification, imageanalysis andunderstandingandsoon.Itis widelyuseinthefieldofindustry, kinesiologyandbiomedicineetc.Thefrequently-usedmethods ofCurveidentification suchas Houghtransform,RandHoughtransform,basedonArccombinationmethodandtheanalysis ofcluster,etc.Inthepaper,theissueofcurveidentificationaboutmulti-circleor arcwillbesolvedwith themethodofcluster.Firstofall,theanalysispointatthedrawback aboutthelarge computingcapacityandtheslowconvergence rateinthetraditionalGA-FCSS,andimprove theselectinitialpopulationintheGAstep.Thearticleisproposeanewmethodtoselectedtheinitial population whichisusingthemethod of “threepointsdefineacircle”.Theexperimentsshowthatthenewmethod has bettereffectsin thecomputingcapacityandtheconvergenceratethanthe traditional GA-FCSSintheclusteranalysis issuesofconcentric circle.Asa resultoftheFCSSis amethodofcurveidentificationbasedonFCM,itsucceedsthedrawbackofFCM suchas sensitiveabouttheinitialvalueandnosepoint,the resulting membershipcan’trepresentthetypicalofclassificationinthedatapoint,etc.Forthis reason, thearticleanalysis theclusteralgorithmFCM-PCM,andintroducethealgebraicdistanceinto thealgorithm.Thereproposeacircleor arcidentificationmethodIPCSSbasedonthemodel ofFCM-PCM.Thesimulateexperiments showthatthe IPCSSalgorithmhas abettercluster resultinthedatapointofconcentriccircletype, anditcansolvethe issuesofsensitiveaboutthe initialvalueandnosepointtoacertainextent.KEYWORDS:Cluster,Fuzzyspherical shell clustering, Geneticalgorithm,fuzzyc-means clustering, Possibilityclustering目录目录.....................................................1第一章绪论...............................................11.1课题研究背景........................................11.2国内外研究现状 .....................错误!未定义书签。1.2.1曲线识别的发展研究现状.......错误!未定义书签。1.2.2基于智能优化算法的聚类方法研究现状.错误!未定义书签。1.3本文组织结构........................................5第二章改进的模糊球壳聚类算法...