一种全自动三维点云配准及比例约束方法  

An Automatic Method of 3D Point Cloud Registration and Dimension Adjustment

作　　者：谢晓燕[1] 吴锦桥 

机构地区：[1]西安邮电大学计算机学院,陕西西安710121

出　　处：《计算机技术与发展》2015年第3期63-66,70,共5页Computer Technology and Development

基　　金：陕西省教育科研计划资助项目(12JK0938)

摘　　要：为了能够对初始相对位置不确定,且尺寸互不一致的多块三维点云进行配准,并对配准后的点云进行尺寸调整,提出了一种新的全自动点云配准及比例约束方法。利用三维重建过程中的相机参数信息,先匹配出特征同名点对,再解算旋转变换矩阵,最后依据摄影测量理论中的共线方程,采用空间后方交会方法求解出点云的真实尺寸。实验结果表明,所提方法能较好地实现初始位置不确定的多块点云之间的配准及三维尺寸调整;另外,相对于最近迭代点算法,新方法实现简单,能获得较好的配准精度。To register the multi-blocks of 3D point cloud which its initial relative position is uncertain and the dimension is different by each other and adjust their dimension,propose a fully new automatic method of 3D point cloud registration and dimension adjustment.First search the feature same-name point pairs by using the camera parameter information produced in the process of three-dimension reconstruction,and then compute rotation matrix based on same-name point pairs. Finally,calculate the point cloud real dimension according to the theory of classical photogrammetry and space resection. Experimental results show that the method can accurately accomplish the registration and dimension adjustment of multi-block point cloud which its initial relative position is uncertain. In addition,the new method proposed can also be implemented easily and get good performance at the aspect of accuracy.

关 键 词：点云 点云配准 比例约束 三维重建 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]