點(diǎn)云配準(zhǔn)方法課件

ABriefIntroductionOfPointCloudRegistrationMethod

點(diǎn)云配準(zhǔn)方法

PointCloudRegistrationMethodPointCloudRegistrationWithTargetControl34QuaternionMethod21IterativeClosestPoint(ICP)Algorithm2點(diǎn)云配準(zhǔn)方法

OutlineTwopointcloudsetsofAandB,areconformationsofthesamepointsindifferentcoordinatesystems

Rigidtransformation:

Objectivefunction:3點(diǎn)云配準(zhǔn)方法1.PointCloudRegistrationWithTargetControl4面

（a）Leicastandardrotationplanetarget（b）MensiplanetargetFigure1Differentkindsofplanetarget（d）Mensiplanetarget（c）FAROstandardplanetarget點(diǎn)云配準(zhǔn)方法1.PointCloudRegistrationWithTargetControl6spatialsimilaritytransformationparameters3angleelements:3translation

elements:

Adjustmentmodel:Or5點(diǎn)云配準(zhǔn)方法1.PointCloudRegistrationWithTargetControl

Rotationmatrix:

6點(diǎn)云配準(zhǔn)方法1.PointCloudRegistrationWithTargetControl

TheexpansionofTaylor'sformula:

ErrorEquation:

7點(diǎn)云配準(zhǔn)方法

G1.PointCloudRegistrationWithTargetControl

Derivativevalue:

Supposethat:8點(diǎn)云配準(zhǔn)方法1.PointCloudRegistrationWithTargetControl

Errorequationmatrix:

Gravity-centralize:9點(diǎn)云配準(zhǔn)方法1.PointCloudRegistrationWithTargetControl

Npairsofcorrespondingpoints:

Order:10點(diǎn)云配準(zhǔn)方法1.PointCloudRegistrationWithTargetControl

Errorequation:

11點(diǎn)云配準(zhǔn)方法1.PointCloudRegistrationWithTargetControl

Newapproximatevalue:

Iteration:Finally,wegetthevaluesofparameter:12點(diǎn)云配準(zhǔn)方法2.IterativeClosestPoint(ICP)Algorithm13Basicthought:Firstly,fromapointset,abarcurve,orasurfacetofindtheclosestpointthatcorrespondstoonepoint,andthenusesthisresulttofindtwocorrespondingpointsets.Finally,wefindoutthecorrespondingpointsetandthecorrespondingtransformationmatrix.Basicsteps:Step1:SearchforthenearestpointsStep2:SolvetransformationrelationsStep3:ApplicationtransformationStep4:Repeatediteration點(diǎn)云配準(zhǔn)方法3QuaternionMethod

WhatisQuaternion?

14Unitquaternions

RotationmatrixR點(diǎn)云配準(zhǔn)方法3.QuaternionMethodDefinitetranslationvector:

15Completeregistrationstation:

Assumethat:

Minimumobjectivefunction:Thecentroidofthesetofbasicpoints:點(diǎn)云配準(zhǔn)方法3.QuaternionMethod16Orthogonalcovariancematrixofpointset:Antisymmetricmatrix’scycliccomponents:Constructasymmetricmatrix:點(diǎn)云配準(zhǔn)方法3.QuaternionMethod17Translationmatrix:Theuniteigenvectorcorrespondingtothema