含有點狀運動目標圖像序列中的背景雜波抑制技術研究.pdf

1、碩士研究生學位論文新疆大學論文題目(中文)：含有點狀運動目標圖像序列中的背景雜波抑制技術研究論文題目(英文)：baS趣roundc1uttersuppressiontechniquesIllimagesequencesWltl!movingpointtargets研究生姓名：學科、專業(yè)：研究方向：導師姓名職稱：吳劍英通信與信息系統(tǒng)H‘多媒體通信與信息處理技術艾斯卡爾教授論文答辯臼期：2006年月日學位授予日期：2006年月日Abstra

2、ctAsweknow舀oalmotioncausedbycameravibrationisquitecommoninimagesequenceThisnuisancefactorresuktsintranslational，totationalandparallaxdistortionsinimagesSo，theimageregistrationprocesshasbecomeacriticalandnonseparableparto

3、fmotionanalysisandsegmemationGlobalmotioncanbemodeledbyafewparameterswillbestatedinchapter2Inchaptertwoaregiventhesystemstructureandkeytechnologiesofglobalestimationandcompensation，describingseveralpossibledetailsofeachk

4、eytechnologyGlobalmotionparametersestimatedusingregressiontechniquewhichfirstestimatethelocalmotionandthenusesthelocalinformationtofindtheglobalmotionthatminimizetheleastsquareellorThentheglobalmotionwascompensatedusingt

5、heseparametersRefinementsarecarriedoutduringtheestimationinnerprocessbasedoniterativeeliminatiohofsingularvalueswhichintroducethebiastotheestimatedparametersTheproposedtechniqueisnotonlycomputationallysimple，butalsohashi

6、ghperformanceTheoreticalanalysisofthetechnique’SperformanceandexperimentalresultsarealsogiveninthispaperAnditisaninevitablethingthatbackgroundclutterisanapplianceoftoimagersensor，Soitimportsmuchthatanapproachforcluttersu

7、ppressiontOconstructthevirtualobserverNonparamctricregressionstudiesoptimalmethodsoffunction(backgroundclutter)estimationfromnoisyobservationsThedistinguishingfeatureofnonparametricregressionisthatthereisno(orverylittle)