基于公共标志点的三维点云拼接方法研究

Research on the Stitching Method for 3D Point Cloud Based on Public Marked Point

针对目前在处理公共标志点配准所面临的算法复杂性高、自动化程度低等问题,提出了一种公共标志点自动配准方法。首先获取基准点对,通过比对各个标志点到基准点的欧式距离来匹配标志点;再利用罗德里格参数进行变换矩阵求解,实现两组点云的粗拼接。为避免最近点迭代算法陷入局部最优,采用粗拼接后两组点云的交集作为初始位置。通过在不同视角下进行拼接实验和精确拼接的误差可以看出,其明显优越于传统的拼接方法。结果表明,该方法快速有效,拼接精度高。

To solve the problem of the high complexity of arithmetic and low degree of automation in handling match-ing of public marked point, in this paper, an automatic matching method for public marked point was put forward. Firstly, a reference point pair was got and marked point was matched by comparing the Euclidean distance of each marked point to reference points. Then,transform matrix solution was got by using the Rodrigues parameter,to realize coarse stitching of two groups of point cloud. Finally,in order to avoid the local optimum of ICP,intersection of two sets of point cloud as the initial position was adopted.This method is obviously superior to the conventional stitching method, that could be seen from the performance of stitching experiments at different viewing angles and the error in the precise stitching.The experiment shows that it is not only fast but also highly stitching precision.