基于中值滤波和小波变换的图像去噪研究-电路与系统专业论文.docx

波与基于高斯混合模型的小波去噪法相结合的图像去噪算法。仿真实验表明，对于被高斯、脉冲混合噪声污染的图像，该算法的去噪效果显然比单一的中值滤波和小波去噪法好得多。关键词：图像去噪；中值滤波；小波去噪；混合噪声IIABSTRACTImagesaredisturbedinevitablybyallkindsofnoisewhencollected andtransmitted．Inordertohigherlevelprocessing，it’Sverynecessarytocarryondenoisingpretreatmenttoimages．Therearemanykindsofnoiseinimages，mostlyareimpulseandGaussiannoise．Medianfilterisanonlinearfilteringtechniquewidelyusedforremovelofimpulsenoise．Wavelettransformisastrongmathematicalanalysist001．Itiswidelyappliedtothefieldofimage denoisinginnearyears．Andmanygoodwaveletimagedenoisingalgorithmshaveappeared，ontheassumptionthatnoiseofimagesisGaussianwhitenoise．ThispaperlucubratedthefilteringofimpulsenoiseandGaussiannoise，whicharethemostcommoninimages．Theresearchcontentsconsistofthreeaspects：(1)Weanalyzedandstudiedstandardmedianfilteranditsimprovementalgorithms．In ordertoovercomeinherentshortcomingofstandardmedianfilter，thepapergivesanadaptivemedianfilteralgorithmbasedondetectionofimpulsenoisepoints，andcomparesitwithotheralgorithmsbysimulationexperiments．Fromtheexperimentalresults，wecanknowthatcomparedwithothermedianfilter,thenewalgorithmcanobtainthebesteffectwhenfilteringimpulsenoise．(2)Wedidsomeresearchworkaboutimagedenoisingmethodsbasedonwavelet．Afteranalyzingwaveletcoefficientcharacteristicofnaturalimages，awaveletimagedenoisingalgorithmbasedonGaussianMixtureModelisgiven．Theexperimentalresults indicatethatthealgorithmis betterthanotherwaveletdenoisingalgorithmswhenremovingGaussianwhitenoise．(3)Usually,one denoisingmethodisonlyeffectivetoonekindofnoise．Inordertosuppressmixednoise，thepapergivesanimagedenoising algorithm,which combinestheadaptivemedianfilterbasedondetectionofimpulsenoisepointswiththewaveletdenoisingbasedonGaussianMixtureModel．SimulationexperimentsshowthatthealgorithmisobviouslymuchbetterthansinglemedianfilterorwaveletdenoisingmethodwhenfilteringmixednoiseofGaussianandimpulseinimages．KeyWords：imagedenoising；medianfilter；waveletdenoising；mixedn01SeIV湖南师范大学学位论文原创性声明本人郑重声明：所呈交的学位论文，是本人在导师的指