《点云配准方法》PPT课件.ppt

《《点云配准方法》PPT课件.ppt》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、ABriefIntroductionOfPointCloudRegistrationMethod2PointCloudRegistrationMethodPointCloudRegistrationWithTargetControl34QuaternionMethod21IterativeClosestPoint(ICP)Algorithm4OutlineTwopointcloudsetsofAandB,areconformationsofthesamepointsindifferentcoordin

2、atesystemsRigidtransformation:Objectivefunction:1.PointCloudRegistrationWithTargetControl4面（a）Leicastandardrotationplanetarget（b）MensiplanetargetFigure1Differentkindsofplanetarget（d）Mensiplanetarget（c）FAROstandardplanetarget51.PointCloudRegistrationWith

3、TargetControl6spatialsimilaritytransformationparameters3angleelements:3translationelements:Adjustmentmodel:Or61.PointCloudRegistrationWithTargetControlRotationmatrix:71.PointCloudRegistrationWithTargetControlTheexpansionofTaylor'sformula:ErrorEquation:8

4、G1.PointCloudRegistrationWithTargetControlDerivativevalue:Supposethat:91.PointCloudRegistrationWithTargetControlErrorequationmatrix:Gravity-centralize:101.PointCloudRegistrationWithTargetControlNpairsofcorrespondingpoints:Order:111.PointCloudRegistratio

5、nWithTargetControlErrorequation:1.PointCloudRegistrationWithTargetControl12Newapproximatevalue:Iteration:Finally,wegetthevaluesofparameter:2.IterativeClosestPoint(ICP)Algorithm13Basicthought:Firstly,fromapointset,abarcurve,orasurfacetofindtheclosestpoin

6、tthatcorrespondstoonepoint,andthenusesthisresulttofindtwocorrespondingpointsets.Finally,wefindoutthecorrespondingpointsetandthecorrespondingtransformationmatrix.Basicsteps:Step1:SearchforthenearestpointsStep2:SolvetransformationrelationsStep3:Applicatio

7、ntransformationStep4:Repeatediteration3QuaternionMethodWhatisQuaternion?14UnitquaternionsRotationmatrixR3.QuaternionMethodDefinitetranslationvector:15Completeregistrationstation:Assumethat:Minimumobjectivefunction:Thecentroidofthesetofbasicpoints:3.Quat

8、ernionMethod16Orthogonalcovariancematrixofpointset:Antisymmetricmatrix’scycliccomponents:Constructasymmetricmatrix:3.QuaternionMethod17Translationmatrix:Theuniteigenvectorcorrespondingtothemaximumeigenvalueofmatrix:ConclusionsTha