资源描述:
《DIP_Lecture_7》由会员上传分享,免费在线阅读,更多相关内容在行业资料-天天文库。
1、ShanghaiJiaoTongUniversityInstituteofImageCommunication&InformationProcessingSchoolofElectronicInformation&ElectricalEngineeringDept.ofElectronicEngineering2010DigitalImageProcessingProject(2)°Description-Compareandanalyzetheperformanceofdifferentimagedenoisingalgorithms°Content-Chooseagrayscaleima
2、ge-Addsalt-and-peppernoisetotheimage-AddGaussiannoisetotheimage-Useatleast2methodstoreducethenoise-Analyzetheexperimentalresults&drawaconclusion-Writeaprojectreportandsendittozhangcong0929@sjtu.edu.cntogetherwithMATLABsourcecodebeforenextThursday2010ShanghaiJiaoTongUniversityInstituteofImageCommuni
3、cation&InformationProcessingChapter5SchoolofElectronicInformation&ElectricalEngineeringDept.ofElectronicEngineering2010DigitalImageProcessing•Goalofimagerestoration-toimprovethequalityofanimage•Imagerestorationattemptstoreconstructorrecoveranimagethathasbeendegradedbyusingaprioriknowledgeofthedegra
4、dationphenomenon•Imageenhancement&imagerestoration-imageenhancementislargelyasubjectiveprocess-imagerestorationisforthemostpartanobjectiveprocess•Imagerestorationtechniquesareorientedtowardmodelingthedegradationandapplyingtheinverseprocessinordertorecovertheoriginalimage•Imageenhancementtechniquesb
5、asicallyareheuristicproceduresdesignedtomanipulateanimageinordertotakeadvantageofthepsychophysicalaspectsofthehumanvisualsystem2010DigitalImageProcessingA.ModeloftheImageDegradation/RestorationProcessA.ModeloftheImageDegradation/RestorationProcess•Thedegradationprocessismodeledasadegradationfunctio
6、nthat,togetherwithanadditivenoiseterm,operatesonaninputimagef(,)xytoproduceadegradedimageg(,)xy•Theobjectiveofimagerestorationistoobtainanestimatefˆ(,)xyoftheoriginalinputimage2010DigitalImageProcessing•Ifisalinear,position-invariantHprocess,thenthedegradedimageisgiveninthespatialdomainbyg(,)xyhxyf
7、xynxy=(,)(,)(,)∗+nx(,)yf(,)xyhx(,)yg(,)xy•InthefrequencydomainGuv(,)=HuvFuvNuv(,)(,)+(,)Nuv(,)F(,)uvHuv(,)Guv(,)2010DigitalImageProcessingB.Linear,PositionB.Linear,Position--InvariantDegradationsInvariantDe