基于分?jǐn)?shù)階次信號處理的新型邊緣檢測方法研究.pdf

1、南京航空航天大學(xué)碩士學(xué)位論文基于分?jǐn)?shù)階次信號處理的新型邊緣檢測方法研究姓名：楊海波申請學(xué)位級別：碩士專業(yè)：控制理論與控制工程指導(dǎo)教師：葉永強201103基于分?jǐn)?shù)階次信號處理的新型邊緣檢測方法研究ABSTRACTEdgeisoneofthemostessentialfeaturesofimage.Edgedetectionistheelementaryfundamentalstageofimageprocessingimageproces

2、singapplicationshavebeenintegratedwithvariousareasofhumanactivity.Howtobalancetheedgedetectionaccuracynoiseimmunityhasalwaysbeenimptantdifficultfallresearchers.Thisthesisisdevotedtodevelopanovelschemetosolvethisconflicta

3、chievesremarkableresults.OnthebasisofanalysissynthesisoftheexistingresearchresultsthethesisproposesannoveledgedetectionoperatYE&YANGoperatwhichhastwoversionsi.e.firstdersecondderversion.Themaincontributionofthethesisarel

4、istedasfollows:Firstlythethesisadoptsfractionaldifferentiationintegrationtoobtainsocalledcompositefirstsecondderderivatives.Thentheamplitudephaseacteristicsofthecompositederivativesareanalyzedtheadvantagesofthecomposited

5、erivativesareillustrated.Theoperatisfirstlyrealizedbythefrequencydomainapproximationmethodthenitspracticalimplementationinedgedetectionisdiscussedfourschemesareproposed.Onebestschemeisedthroughexperiments.Thenthefraction

6、aldersignalprocessingbasedfirstderedgedetectionmethodfractionalderssignalprocessingbasedsecondderedgedetectionmethodareproposed.Theimplementationprocessesofthetwomethodstogetherwiththeeticalstudyalgithmanalysisparameterd

7、esignareelabatedindetails.Thedetectionaccuracynoiseimmunityofthetwomethodsareillustrated.Theadvantagesareanalyzedhighlighted.FinallythedetectionresultsofthetwomethodsarecomparedwiththosebytheclassicalalgorithmssuchasCANN

8、YoperatLOGoperatetc.usingbothsubjectiveobjectiveevaluationcriterion.Theeffectivenessofthemethodsareproved.Theexperimentalresultsverifythatthenewoperatispromisingalthoughtherearedeficienciesthatneedtobefurtherstudied.Keyw