您好,欢迎访问三七文档
当前位置:首页 > 商业/管理/HR > 公司方案 > 利用深度图像进行场景的三维重建
100871e-mail:{wenfeng,zha}@cis.pku.edu.cn3D:03D3D1234513D3DI.StamosP.Allen[1]S.F.El-Hakim[2]Y.Yu[3]H.,Zhao[4]C.Frueh[5]3DBeslJain[6]I.StamosP.Allen[1]BeslMcKayICPIterativeClosestPointI.StamosP.Allen[1]Y.Yu[3]1[8][9]G.Turk[10]2))()((1∑=−⋅−NiiTimvmvCyraTechonogies,Inc.[8]Cyrax2500Cyclone3.13D(x,y,z)[1]Pk*kA=mAPP12iv3313x3R3tR,tRtRtPP’P”PP’P”PRt32n,mnmkkiniiimiimn′′2[01]iwRˆ−′it=R∑=−′kiiiiRnnw12[9]tt∑=′kiTiimnw1((()(1kkiTiiinnw′′∑∑==+itmR2)))ˆ)ˆ(1iiTiiimRmnnw−′′′i33331kRtkRtk=1k=2k=3R,t1k=12k=2332R3.3.133333R3.23334440127AGFEDCBCenterofLaserscanner41mxn(m-1)x(n-1)34747G.Turk[10]A,…,GA,B,C,DE,F,G334552a)b)341676[1]I.Stamos,P.E.Allen,3-Dmodelconstructionusingrangeandimagedata.ComputerVisionandPatternRecognition,HiltonHeadIsland,2000,p.531-6[2]S.F.El-Hakim,C.Brenner,G.Roth,Amulti-sensorapproachtocreatingaccuratevirtualenvironments,JSPRSJournalofPhotogrammetry&RemoteSensing,1998[3]Y.Yu,A.Ferencz,J.Malik,ExtractingObjectsfromRangeandRadianceImages,IEEETrans,VisualizationandComputerGraphics,vol.7.,2001[4]H.Zhao,R.Shibasaki,Asystemforreconstructingurban3Dobjectsusingground-basedrangeandCCDsensors.InUrbanMulti-Media/3DMappingworkshop,InstituteofIndustrialScience(IIS),TheUniversityofTokyo,1999[5]C.Frueh,A.Zakhor,3Dmodelgenerationofcitiesusingaerialphotographsandgroundlevellaserscans,ComputerVisionandPatternRecognition,Hawaii,USA,2001,p.II-31-8,vol.2.2[6]P.J.Besl,R.C.Jain.Segmentationthroughvariable-ordersurfacefitting.IEEETransactionsonPatternAnalysisandMachineIntelligence,10(2):167-192,march1988.[7]P.J.Besl,N.D.McKay,AMethodforRegistrationof3-DShapes,IEEETrans.PatternAnalysisandMachineIntelligence,vol.18,no.5,pp.239-256,May1992.[8]CyraTechonogies,Inc.Onlinedocuments,[9]----1998[10]G.Turk,M.Levoy,Zipperedpolygonmeshesfromrangeimages,InProceedingsofSIGGRAPH’943DReconstroctionofScenesUsingRangeImagesWenfengHe,HongbinZhaCenterforInformationScience,PekingUniversity,Beijing,100871e-mail:{wenfeng,zha}@cis.pku.edu.cnAbstract:Recently,therehasbeenalargeinterestinthe3Dmodelingofrealscenesusingrangedatascannedbylasersensors.Afteracquisitionofrangedata,lotsofworkshouldbecoveredtocreateafinal3Dmodelofascene,andamongthemwefocusonsegmentation,registrationandmeshinginthispaper.Consideringthatthereoftenexistagreatnumberofplanesinrealscenes,wepresentarangeregistrationmethodbasedonplanarfeatures,whichareextractedwhenwesegmentrangeimages.Theresultsofourexperimentsindicatethatthemethodisaccurateandrobust.Inaddition,weimprovethegeneralmethodofcreatingmeshesfromrangeimages,consequentlywegetbetterresultsofmeshing.Keywords:RangeimageLaserscannerPointcloudSegmentationRegistrationMeshing5a)b)52a)b)c)3()d)6e)6a)b)c)a)d)b)77
本文标题:利用深度图像进行场景的三维重建
链接地址:https://www.777doc.com/doc-3621000 .html