资源描述：
《外文翻译基于emd的红外图像目标检测方法.doc》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。
1、英文翻译系别专业班级学生姓名学号指导教师EMDBasedInfraredImageTargetDetectionMethodHeDeng&JianguoLiu&HongLiAbstractUnderthecomplicatedbackgroundofinfraredimage,thesmalltargetdetectionisavitalchallengingtaskinmodernmilitary.Inordertosolvethisproblem,anovelmethodbasedonthe
2、empiricalmodedecomposition(EMD)isproposedinthepaper,todetectsmalltargetsundercomplicatedsea-skybackground.Thedetectionprocesscontainstwosteps:thefirststepistosuppressthesea-skybackgroundoftheinfraredimagebasedonEMD;thesecondstepistosegmentthetargetfr
3、omthebackgroundsuppressedimagethroughathreshold.Theapplicationofinfraredimageshasshownthattheperformanceofthealgorithmcandetectinfraredsmalltargetundersea-skybackgroundexactly.Comparedwithwavelettransformation,thetestingresultsbasedonEMDmethodachieve
4、tantamountresultswavelettransformation,andevenbetterinsomerespects.ThesimulationsshowthatEMDmethodpresentedinthispaperappearsinstructiveforboththeoreticalandpracticalpointsofview.Keywords:Smalltargetdetection,Infraredimage,Backgroundsuppression,Thres
5、hold,EMD.1IntroductionUnderthecomplicatedbackgroundofinfraredimage,thesmalltargetdetection,identificationandtrackingapplicationsinmodernmilitaryarevitalchallengingtasks.Ithasbeenresearchingformanyyearsoninfraredimagingsystemsandautomatictargetdetecti
6、on[1].Thecomplexityofthisproblemariseswhenthetargetissmall,faintandpartiallyobscuredbysurroundingobjects,embeddedinclutters.Inthesecomplexsituations,thefeaturesofboththetargetandthebackgroundaregenerallynonseparableintheoriginalimagespace.Itisdifficu
7、ltforthosedetectionalgorithmstoworkintheoriginalimagespace.So,theimagehastobetransformedintoaso-calledfeaturespace,inwhichthosefeaturescanbeseparated.Sincetargetsandcluttershavedifferentspatialfrequencycharacteristics,aspatialfiltercouldbedesignedtos
8、uppressanddetecttargets.BhanuandJonessummarizedalotofalgorithmsforautomatictargetdetectioninstaticinfraredimagesthatweredevelopeduptotheearly1990s[2].Thesealgorithmspredominatelyutilizetraditionalimage-processingapproachesforopticalpicturesprocessing
