中文摘要：

点云数据配准过程中，传统的七参数、罗德里格矩阵等算法对特征点匹配精度依赖较大，使其在粗差点的情况下，配准精度和效率均受到较大影响。本文通过对罗德里格坐标变换矩阵本身进行分析，提出了一种基于整体最小二乘的罗德里格矩阵算法。该算法能够顾及系数矩阵存在的误差，减小算法实现过程中特征点坐标误差对参数求解的影响。利用实测的三维点云数据配准实验表明，本文方法较传统算法具有更高的精度及稳健性，在初始对应点坐标误差较大的情况下，仍能获得精度稳定的变换参数。

英文摘要：

Since traditional methods such as the seven parameters and Rodriguez matrix greatly depend on the accuracy of feature point matching in the process of point cloud registration, the gross error has a marked effect on matching accuracy and effi- ciency. Based on the analysis of Rodriguez coordinate transformation matrix, a modified Rodriguez matrix algorithm based on the total least squares method is proposed. Errors in the coefficient matrix are considered, and the influence of coordinate error of fea- ture point on parameter solution can be reduced. Registration experiment is conducted using actual 3D point cloud data, and the results show that the proposed method has higher precision and robustness compared with traditional algorithm, and the accuracy of transformation parameter is more stable when the initial coordinate error is great.