适应性距离函数与迭代最近曲面片精细配准  

Fine registration with adaptive distance function and iterative closest surface

作　　者：张梅[1] 彭星星 

机构地区：[1]贵州财经大学信息学院,贵州贵阳550025

出　　处：《浙江大学学报（工学版）》2017年第10期1920-1927,共8页Journal of Zhejiang University：Engineering Science

基　　金：国家自然科学地区基金资助项目(41261094);贵州省科技厅科学技术基金资助项目(黔科合基础[2016]1020);贵州省教育厅自然科学拔尖人才基金资助项目(黔教合KY字[2016]069)

摘　　要：针对复杂曲面物体多视角激光扫描点云数据,提出从深度图像到完整几何模型的配准方法.根据空间点相对位置在刚体变换下的不变特性,利用曲率不变特征和归一化零均值互相关系数构造有效的初始匹配点对数组.基于单位四元数对匹配的特征点对进行坐标变换求解,完成数据粗略配准.探讨改正系数的确定方法与步骤,计算不同改正系数下的均值误差,得到最佳改正系数.运用适应性距离函数和改进迭代最近曲面片精细匹配技术,将不同视角点云在三维空间进行最优化匹配.根据匹配结果计算配准误差,对配准精度和速度进行统计分析.数值试验结果表明,该方法在保证配准精度的前提下能够有效地提高配准效率.A registration method from three-dimensional depth image to three-dimensional geometric models was proposed aiming at laser scanning point cloud data from multi angle of the complex curved surface object.The effective initial matching points array was constructed with curvature invariant features and normalized zero-mean cross-correlation coefficient（NZCC）according to the invariant characteristics of the relative position of the space points in the condition of rigid body transformation.The coordinate transformation of the matching feature points was solved based on the unit quaternion.Then the data coarse registration was completed.The method and procedure for determining the modified coefficient were discussed,and mean error of different correction factors was calculated in order to get the best modified coefficient.The different perspectives of clouds were optimally matched in three-dimensional space by using fine matching technology based on adaptive distance function and the improved iterative closest surface.The registration error was calculated according to the matching results,and the registration accuracy and speed were analyzed.The numerical experiments results show that the method can effectively improve the efficiency of registration in the guarantee of the registration accuracy.

关 键 词：复杂曲面 数据配准 曲率特征 适应性距离函数 改正系数 迭代最近曲面片 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]