通过引入特征点和改进最近点迭代法, 提出了一种

在三维扫描系统中对三维点云数据进行配准的方法。该方法

通过对特征点的提取, 首先得到一组匹配点对, 然后运用

SVD 矩阵分解算法求出转换参数R 和T, 进而以此作为最

近点迭代法的初始值, 并对最近点的求法和迭代截止条件作

了改进, 得到了很好的配准效果。该文论述了该方法的基本

原理, 并通过不同视觉下物体三维测量点云数据配准的应用

实例证明了该方法的有效性。-A 3-D meas uring dat a r egis trat ion method for 3-D

scann ing syst em was developed based on feature point s an d it erat ive

closet point ( ICP) algorithm. T he feature p oin ts are firs t ident ified

an d mat ched. T hen, the rot at ion mat rix R an d the t rans lat ion vector

T are fou nd usin g the s ingul ar value decomposit ion algorithm.

T hus, init ialized b y the former resul t , the ICP algorithm then leads

to a perfect regis t rat ion aft er improving on fin ding clos et d ist an ce

an d d ecid ing wh en to terminat e th e iterations . T he basic principle of

this method is discuss ed, and a regist rat ion applicat ion example of

3-D meas uring data taken from differen t viewpoint s is carried out to

sh ow the perfect feasibil ity of this approach