Removal of Salt and Pepper Noise

《Removal of Salt and Pepper Noise》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、Dr.M.Anandet.al/InternationalJournalonComputerScienceandEngineering(IJCSE)RemovalofSaltandPepperNoisefromhighlyCorruptedImagesusingMeanDeviationstatisticalparameterDr.M.AnandM.Tech.,Ph.DFacultyofDepartmentofECEDr.M.G.RUniversityChennai,Tamilnadu,Indiaharshi

2、ni.anand@gmail.comY.NarasimhaMurthy1FacultyofDepartmentofElectronicsSSBNCOLLEGE(Autonomous)Anantapurcity,A.P,Indiayayavaram@yahoo.comAbstract—Anovelideaofusingmediandeviationparameterinestimatingthenoiseintheimagesisproposedandsuccessfullyappliedtobothgrayl

3、evelaswellascolorimages.Themedianfilterwhichisverypopularinremovingthesaltandpeppernoisefromtheimages,hasundergonemanychangesinrecentpast.Tothismodifiedmedianfiltertheconceptofmediandeviationisaddedandusedinestimatingandremovingthenoise.Theproposedmethodisi

4、mplementedbydevelopingaGraphicalUserInterfaceinMATLABandalsoimplementedusingtheSpartan3EFiledProgrammableDevice.Theresultsarefoundtobebetterthanearliermethodsandalsorobustintermsofpreservingthecontrastandfinedetailsoftheimageevenathighnoisedensities.Keyword

5、s-Medianfilter;Mediandeviation;SaltandPeppernoise;MATLAB;ImageprocessingI.INTRODUCTIONDuringthelastfewyearstherehavebeeninnumerablenumberofresearchpaperspublishedinvariousjournalsontheapplicationofmedianfiltersforremovalofsaltandpeppernoisefromtheimages,byv

6、ariousauthors[1,2,3].Thesuccessofmedianfilterscanbeattributedtotwointrinsicproperties:edgepreservationandefficientnoiseattenuationwithrobustnessagainstimpulsivetypenoise.Edgepreservationisessentialinimageprocessingduetothenatureofvisualperception.Edgesalsoo

7、ccurinbiomedicalsignalswhenthe“system”movesfromonestatetoanother.Thefactthatsomesignalsareinvarianttomedianfilteringoffersinterestingpossibilities.Innoisefiltering,thebasicadeaishowtopreservesomedesiredsignalfeatureswhileattenuatingthenoise.Anoptimalsituati

8、onwouldariseifthefiltercouldbedesignedsothatthedesiredfeatureswereinvarianttothefilteringoperationandonlynoisewouldbeaffected.Asthesuperpositionprinciplecannotbeappliedtonon-linearfilters,thiscanneverb