資源描述：
《Edge Detection of Image on the Local Feature .pdf》由會員上傳分享，免費在線閱讀，更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫。
1、SecondInternationalSymposiumonIntelligentInformationTechnologyApplicationEdgeDetectionofImageontheLocalFeature122ChunhuaLi,KunHe,JiliuZhou1SchoolofElectronicsAndInformationEngineering,SichuanUniversity,Chengdu,SiChuan,610065,China2ComputerCollage,SichuanUniversity,C
2、hengdu,SiChuan,610065,Chinalchhua111@126.comAbstractrestrainstheGaussiannoiseinsomeextent.ThesedetectionoperatorsareallbasedonlinearedgedetectionThetraditionalmethodsofimageedgedetectionaretheory;howevertheimagesignalsarenonlinear,sousingsensitivetothenoisecomparati
3、velyandcannotlocatethethelinearoperatortodetecttheedgewillleadtothenon-[5]edgeexactly.Aimingattheshortageofthetraditionalintegralityofedgeinformation.Levputsforwardthemethodsofimageedgedetection,thepaperputsforwardweightedmodelandGemanputsforwardtheSimulatedtheedged
4、etectionofthelowsignalnoiseratioimage(theNealingmethodtodetecttheedge,andthesetwomethodslowSNRimage)accordingtothelocalfeature.Firstly,webelongtoadaptiveedgedetection,whosebasicideaisanalyzetherelationshipoftheneighborhoodofpixelsandsmoothingthesignalbyacurrencyoper
5、ator,whichcanintroducethefunctionexpressionwhichisusedtomakeitselfsuittothelocalstructureofthesignal,sothisestimatewhetherthepixelisontheimageedgeornot.methodcanholdthemutationaledgeinformationwell.Themethodtakesthepixelastheresearchobject,andtheButforthenon-mutatio
6、naledgetheadaptiveedgeedgetobedetectedisperhapsthenoiseorjustitself.detectionalgorithmisinvalidation.Accordingtotheedge’scontinuityoftheimageandtheThemathematicalmorphologyisthetheoryofthe[6]isolationofnoise,weusethemorphologytoextractthenonlinearimageprocessinganda
7、nalysisanditspurnsimageedgeandovercomethenoiseeffectontheedge,thetraditionalnumericalvaluemodelingandthethenusethemorphologicalgradienttoeliminatethefalseanalyticalmethodintheedgedetection,anddepictsandedgepoints.Withthenewmethod,theimageedgecanbeanalysestheimagefro
8、mtheperspectiveoftheset.detectedmorecorrectly.Moreoverthismethodcandistinguishthenoiseandtheedgefromthemorphology,andcandetecttheedgeof1.I
