图像空间与手术空间点配准算法  

Surgery space and real space registration based on points

作　　者：周振环[1] 徐人凤[1] 

机构地区：[1]深圳职业技术学院软件工程系,广东深圳518055

出　　处：《计算机工程与应用》2007年第6期233-235,共3页Computer Engineering and Applications

基　　金：深圳市科技计划项目(No.05KJcd004);2005年粤港关键领域重点突破项目(No.2005A11304003)。

摘　　要：手术导航系统中手术空间与图像空间点配准方法,普遍使用无约束最小二乘法求出三维刚体变换矩阵,由于测量误差,求出的矩阵存在着缩放、错切和反射等形变,不是严格的三维刚体变换矩阵。提出有约束最小二乘法,约束条件为正交矩阵,使用迭代法使变换矩阵正交化,从而逼近最优变换矩阵,从理论上解决了点配准算法的工程实现。In the surgery navigation system,registrations between surgery space and image space are based on paints matching, unconstrained Least-Squares is usually used for 3D rigid transform matrix,but the matrix exists deformation of shear,scale and reflect,is not strict orthogonal matrix because of measurement error.This paper presents constrained Least-Squares,constraint condition is the orthogonal matrix,and iteration is used to make the matrix orthogonal,up to optimization matrix.Practical implementation of point matching can be solved theoretically.

关 键 词：图像配准 点匹配 手术导航系统 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]