节点文献

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

  • 推荐 CAJ下载
  • PDF下载
  • 不支持迅雷等下载工具,请取消加速工具后下载。

【作者】 杨帆白宝兴张振普刘文静

【Author】 YANG Fan;BAI Baoxing;ZHANG Zhenpu;LIU Wenjing;School of Computer Science and Technology,Changchun University of Science and Technology;

【机构】 长春理工大学计算机科学技术学院

【摘要】 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。

【Abstract】 To solve the problem of the high complexity of arithmetic and low degree of automation in handling matching 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.

【关键词】 计算机视觉拼接标志点自动配准
【Key words】 computer visionstitchingmarked pointautomatic matching
【基金】 吉林省重点科技成果转化项目(20130303011GX)
  • 【文献出处】 长春理工大学学报(自然科学版) ,Journal of Changchun University of Science and Technology(Natural Science Edition) , 编辑部邮箱 ,2014年02期
  • 【分类号】TP391.41
  • 【被引频次】21
  • 【下载频次】327
节点文献中: 

本文链接的文献网络图示:

本文的引文网络