資源描述：
《Edge Feature Extraction Based on Digital Image》由會員上傳分享，免費(fèi)在線閱讀，更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫。
1、ProceedingsoftheIEEEInternationalConferenceonAutomationandLogisticsQingdao,ChinaSeptember2008EdgeFeatureExtractionBasedonDigitalImageProcessingTechniquesFeng-yingCuiandLi-junZouBeiSongCollegeofAutomationandElectricalEngineeringCollegeofAutomationandElectricalEngineeringQingdaoUn
2、iversityofScienceandTechnology,QingdaoUniversityofScienceandTechnology,Qingdao266042,ChinaQingdao,ShanDongProvince,China{cfy_angel&zlj105983}@163.compipisongbei@163.com?Abstract-Edgedetectionisabasicandimportantsubjectincomputervisionandimageprocessing.InthispaperwediscussII.IMA
3、GEDENOISINGseveraldigitalimageprocessingtechniquesappliedinedgeAsweallknow,theactualgatheredimagescontainfeatureextraction.Firstly,wavelettransformisusedtoremovenoisesintheprocessofformation,transmission,receptionandnoisesfromtheimagecollected.Secondly,someedgedetectionprocessin
4、g.Noisesdeterioratethequalityoftheimage.TheyoperatorssuchasDifferentialedgedetection,Logedgedetection,CannyedgedetectionandBinarymorphologyareanalyzed.Andmakeimageblur.Andmanyimportantfeaturesarecoveredup.thenaccordingtothesimulationresults,theadvantagesandThisbringslotsofdiffic
5、ultiestotheanalysis.Therefore,thedisadvantagesoftheseedgedetectionoperatorsarecompared.ItmainpurposeistoremovenoisesoftheimageinthestageofisshownthattheBinarymorphologyoperatorcanobtainbetterpretreatment.edgefeature.Finally,inordertogainclearandintegralimageThetraditionaldenoisi
6、ngmethodistheuseofalow-passprofile,themethodofborderingclosedisgiven.Afterorband-passfiltertodenoise.Itsshortcomingisthatthesignalexperimentation,edgedetectionmethodproposedinthispaperisisblurredwhennoisesareremoved.Thereisirreconcilablefeasible.contradictionbetweenremovingnoise
7、andedgemaintenance.YetwaveletanalysishasbeenprovedtobeapowerfultoolforIndexTerms-Edgedetection,digitalimageprocessing,imageprocessing[2].BecauseWaveletdenoisingusesaoperator,waveletanalysisdifferentfrequencyband-passfiltersonthesignalfiltering.ItI.INTRODUCTIONremovesthecoefficie
8、ntsofsomescaleswhichmainlyreflectTheedgeisaseto
