图像识别外文文献翻译.docx

图像识别外文文献翻译.docx

《图像识别外文文献翻译.docx》由会员分享，可在线阅读，更多相关《图像识别外文文献翻译.docx（27页珍藏版）》请在冰点文库上搜索。

图像识别外文文献翻译

译文

学院电气信息工程学院

专业计算机科学与技术

学生姓名XXX

班级学号1045532111

指导教师XXX

二零一四年六月

DigitalImageProcessingandEdgeDetection

TurgayCelikHasanDemire1,HuseyinOzkaramanli,MustafaUyguroglu

LDigitalImageProcessing

Interestindigitalimageprocessingmethodsstemsfromtwoprincipalapplicantionareas:

improvementofpictorialinformationforhumaninterpretation;andprocessingofimagedataforstorage,transmission,andrepresentationforau・tenuousmachineperception.

Animagemaybedefinedasatwo-dimensionalfunction,Rx,y）,wherexandyarespatial（planc）coordinates,andtheamplitudeoffatanypairofcoordinates（x,y）iscalledtheintensityorgrayleveloftheimageatthatpoint・Whenx,y,andtheamplitudevaluesoffarcallfinitc,discretequantities,wccalltheimageadigitalimage・Thefieldofdigitalimageprocessingreferstoprocessingdigitalimagesbymeansofadigitalcompute匚Notethatadigitalimageisconposedofafinitenumberofelements,eachofwhichhasaparticularlocationandvalue.Theseelementsarereferredtoaspictureelements,imageelements,peels,aixipixels・Pixelisthetermmostwidelyusedtodenotetheelementsofadigitalimage・

Visionisthemostadvancedofoursenses,soitisnotsurprisingthatimagesplaythesinglemostimportantroleinhumanperception・However,unlikehumans,whoarelimitedlothevisualbandoftheelectromagnetic（EM）spec-trump,imagingmachinescoveralmosttheentireEMspectrunirangingfromgammatoradiowaves.Theycanoperateonimagesgeneratedbysourcesthathumansarenotaccustomedtoassociatingwithimages.Theseincludeultra-sound,electronmicroscopy,andcomputer-generatedimages・Thus,digitalimageprocessingencompassesawideandvariedfieldofapplications・

Thereisnogeneralagreementamongauthorsregardingwhereimageprocessingstopsandotherrelatedareas,suchasimageanalysisandcomputervi-son,start.Sometimesadistinctionismadebydefiningimageprocessingasadisciplineinwhichboththeinputandoutputofaprocessareimages・Webelievethistobealimitingandsomewhatartificialboundary.Forexample,uixlerthisdefinition,eventhetrivialtaskofcomputingtheaverageintensityofanimage（whichyieldsasinglenumber）wouldnotbeconsideredanimageprocessingoperationOntheotherhand^therearefieldssuchascomputervisionwhoseultimategoalistousecomputerstoemulatehumanvision,includinglearningandbeingabletomakeinferencesandtakeactionsbasedonvisualinputs・Thisareaitselfisabranchofartificialintelligence（Al）whoseobjectiveistoemulatehumanintelligence・ThefieldofAlisinitsearlieststagesofinfancyintermsofdevelopment,withprogressleavingbeenmuchslowerthanoriginallyanticipated・Theareaofimageanalysis（alsocalledimageunderstanding）isinbc・teenimageprocessingandcomputervision・

Therearcnoclcar-cutboundariesinthecontinuumfromimageprocessingatoneendtocomputervisionattheother.However,oneusefulparadigmistoconsiderthreetypesofcomputerizedprocessesinthiscontinuum:

low-,mid-,andhigh・levelprocesses.Low-levelprocessesinvolveprimitiveopera・tonssuchasimagepreprocessingtoreducenoise,contrastenhancement,andimagesharpening・Alow-levelprocessischaracterizedbythefeetthatbothitsinputsandoutputsareimages.Mid-levelprocessingonimagesinvolvestaskssuchassegmentation（partitioninganimageintoregionsorobjects）,descriptionofthoseobjectstoreducethemtoaformsuitableforcomputerprocessing,andclassification（recognition）ofindividualobjects.Amidlevelprocessischaracterizedbythefactthatitsinputsgenerallyareimages,butitsoutputsareattributesextractedfromthoseimages（e.g.,edges,contours,andtheidentityofindividualobjects）.Finally,higher-levelprocessinginvolves"makingsense"ofanensembleofrecognizedobjects,asinimageanalysis,and,attheforendofthecontinuum,performingthecognitivefunctionsnormallyassociatedwithvision.

Basedontheprecedingcomments,wcseethatalogicalplaceofoverlapbetweenimageprocessingandimageanalysisistheareaofrecognitionofindividualregionsorobjectsinanimage・Thus,whatwecallinthisbookdigitalimageprocessingencompassesprocesseswhoseinputsandoutputsareimagesand,inaddition9eneompassesprocessesthatextractattributesfromimages,uptoandincludingtherecognitionofindividualobjects.Asasimpleillustrationtoclarifytheseconcepts.considertheareaofautomatedanalysisoftext.Theprocessesofacquiringanimageoftheareacontainingthetext,preprocessingthatimage,extracting（segmenting）theindividualcharacters,describingthecharaclersinaformsuitableforcomputerprocessing,andrecognizingthoseindividualcharactersareinthescopeofwhatwecalldigitalimageprocessinginthisbook.Makingsenseofthecontentofthepagemaybeviewedasbeinginthedomainofimageanalysisandevencomputervisioadepeixiingonthelevelofcomplexityimpliedbythestatement"makingsense."Aswillbecomeevidentshortly,digitalimageprocessing,aswehavedefinedit,isusedsuccessfullyinabroadrangeofareasofexceptionalsocialandeconomicvalue.

Theareasofapplicationofdigitalimageprocessingarcsovariedthatsomeformoforganizationisdesirableinattemptingtocapturethebreadthofthisfield.Oneofthesimplestwaystodevelopabasicunderstandingoftheextentofimageprocessingapplicationsistocategorizeimagesaccordingtotheirsoiucc（c.g・，visual.X-ray,andsoon）.TheprincipalenergysourceforimagesinusetodayistheelectromagneticenergyspectrumOtherimportantsourcesofenergyincludeacoustic,ultrasonic,andelectronic（intheformofelectronbeamsusedinelectronmicroscopy）・Syntheticimages,usedformodelingandvisualization^aregeneratedbycompute匚Inthissectionwediscussbrieflyhowimagesaregeneratedinthesevariouscategoriesandtheareasinwhichtheyareapplied・

ImagesbasedonradiationfromtheEMspectrumarethemostfamiliar,especiallyimagesintheX-rayandvisualbandsofthespectrumElectromagneticewavescanbeconceptualizedaspropagatingsinusoidalwavesofvaryingwavelengtlis,ortheycanbethoughtofasastreamofmasslessparticles,eachtravelinginawavelikepatternandmovingatthespeedoflightEachmasslessparticlecontainsacertainamount（orbundle）ofenergy・Eachbundleofenergyiscalledaphoton.【fspectralbandsarcgroupedaccordingtoenergyperphoton,wcobtainthespectrumshowninfig.below,rangingfromgammarays（highestenergy）atoneendtoradiowaves（lowestenergy）attheother.ThebandsarcshownshadedtoconveythefactthatbandsoftheEMspectrumarcnotdistinctbutrathertransitionsmoothlyfi*omonetotheother.

Imageacquisitionisthefirstprocess・NotethatacquisitioncouldbeassimpleasbeinggivenanimagethatisalreadyindigitalformGenerally,theimageacquisitionstageinvolvespreprocessing,suchasscaling・

Imageenhancementisamongthesimplestandmostappealingareasofdigitalimageprocessing.Basically,theideabehindenhancementtechniquesistobringoutdetailthatisobscured,orsimplytohighlightcertainfeaturesofinterestinanimage.Afamiliarexampleofenhancementiswhenweincreasethecontrastofanimagebecause"itlooksbetter.^Itisimportanttokeepinmindthatenhancementisaverysubjectiveareaofimageprocessing.Imagerestorationisanareathatalsodealswithimprovingtheappearanceofanimage・However,unlikeenhancement,whichissubjective,imagerestorationisobjective,inthesensethatrestorationtechniquestendtobebasedonmathematicalorprobabilisticmodelsofimagedegradation.Enhanccincnt,ontheotherhand,isbasedonhumansubjectivepreferencesregardingwhatconstitutesa“good"enhancementresult・

ColorimageprocessingisanareathatliasbeengaininginimportaneebecauseofthesignificantincreaseintheuseofdigitalimagesovertheInternet.Itcoversanumberoffuixiamentalconceptsincolormodelsandbasiccolorprocessinginadigitaldomain・Colorisusedalsoinlaterchaptersasthebasislorextractingfeaturesofinterestinanimage.

Waveletsarethefoundationforrepresentingimagesinvariousdegreesofresolution.Inparticular,thismaterialisusedinthisbookforimagedatacompressionandforpyramidalrepresentation,inwhichimagesaresubdividedsuccessivelyintosmallerregions.

Compression,asthenameimplies,dealswithtechniquesforreducingthestoragerequiredsavinganimage,orlhebandwidthrequiredtransmittingit.Althoughstoragetechnologyhasimprovedsignificantlyoverthepastdecade,thesamecannotbesaidfortransmissioncapacity・ThisistrueparticularlyinusesoftheInternet,whicharecharacterizedbysignificantpictorialcontent.Imagecompressionisfemiliar（perhapsinadvertently）tomostusersofcomputersintheformofimagefileextensions,suchasthejpgfileextensionusedintheJPEG（JointPhotographicExpertsGroup）imagecomprcssionstandard・

Moipliologicalprocessingdealswithtoolsforextractingimagecomponentsthatarcusefulintherepresentationanddescriptionofsliape.Thematerialinthischapterbeginsatransitionfromprocessesthatoutputimagestoprocessesthatoutputimageattributes.

Segmentationprocedurespartitionanimageintoitsconstituentpartsorobjects.Ingeneral,autonomoussegmentationisoneofthemostdifficulttasksindigitalimageprocessing.Aruggedsegmentationprocedurebringstheprocessalongwaytowardsuccessfulsolutionofimagingproblemsthatrequireobjectstobeidentifiedindividually