基于特征提取的sar图像去斑方法

《基于特征提取的sar图像去斑方法》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、AbstractAbstractSyntheticapertureradar(SAR)isakindofhighresolutionimagingradar.Ithastheabilityoftakingimagesundernearlyallweatherconditionsandimaginganytargetontheearth,soithasbeenwidelyappliedtomilitaryreconnaissancefieldsandnationaleconomicfields.Inrecentyears，withtherapiddeve

2、lopmentofSARsatelliteastronomytechnique，SARimageprocessingisbecomingmoreandmoreimportant．Unfortunately,SARimagesaremuchdegradedbyspecklebecauseofthecoherentimagingsystem.Speckleaffectsobjectdetectionandidentification.Therefore,itisofgrestimportancetosuppressthespeckleinSARimages

3、.Thereisadilemmacontradictionsbetweenedgeinformationandnoisesuppressionprotectioninthetraditionalmethodofimagedenoising.ThisthesisisfocusedontheSARimagedespecklingbasedontargetsextraction.(1)ThisthesispresentsaSARimagespecklefilteringcombiningtargetextractionwithSRADalgorithm.Th

4、emethodfirstdetecttargets,suchasisolatedpoints,linesandtexturefeatures,bybilateralfilteringandCannyedgedetectionoperator,andthenextractandremovethemfromtheSARimage.ThenimplySRADfileringwiththeoptimumnumberofiterationstothetargets-removedimage.Thismethodcanremovethemostspecklenoi

5、seinimagesandretainasmuchaspossibleofthedetailsoftheimageatthesametime.(2)ThisthesisfirstanalysistheBM3Dalgorithm,andthenproposedaSARimagespecklefilteringcombiningtargetextractionwithBM3Dalgorithm.ThealgorithmfirstextractsthetargetimagesandremovethemfromtheSARimage.ThenapplyBM3D

6、filteringtothetargets-removedimage.ThismethodintroducestheBM3Dalgorithmwhichisamoreeffectivedenoisingmethod.TosolvetheproblemoftheSRADalgorithm,whichistoosimpletoachievethebestdespeckledresults.Theexperimentsshowthatthemethodcandespecklefullyandpreservepointtargets，edgesandtextu

7、rewell．(3)ThisthesisproposedaSARimagedespecklingmethodbasedontheK-meansclusteringandPPBalgorithm.ThemethodextractthedetailtargetsoftheimageusingK-meansclustering.Thendespecklethetargets-removedimageusingPPBalgorithm.Thismethodcanpreserveisolatedpointtargetsbetterthanothermethods

8、.Keyword:Syntheticapertureradarspecklenoisefeat