摘要
在已有相关理论的基础上,研究了ICP算法中的采样策略及匹配点集的确定。首先确定重叠区域,识别边界点;其次对重叠区域进行采样,在ICP算法的迭代过程中的不同阶段采取不同的采样策略,同时提出一种基于法矢三点插值的方法确定匹配点,保证配准的精度,提高配准的速度;最后通过对点云数据的测试,验证了研究成果的有效性和实用性。
On the basis of existing related theory, this paper placed emphasis on sampling strategy and determination of the matching point set of ICP algorithm. Firstly,the overlap area was determined and the point on the boundary was recognized. Secondly, sampling in the overlap area, and took differ- ent sampling strategies in different stages of the iteration of ICP algorithm. At the same time,this pa- per presented a new way to determine the match point set to guarantee the precision and increase the speed of the fine registration. Finally, experiments based on the point data verified the effectiveness and practicability of the algorithm.
出处
《中国机械工程》
EI
CAS
CSCD
北大核心
2014年第4期497-502,共6页
China Mechanical Engineering
基金
"十二五"国家科技支撑计划资助项目(2012BAI07B04)
国家自然科学基金资助项目(51205192)
关键词
精确配准
采样策略
协方差矩阵
匹配点集
fine registration
sampling strategy
covariance matrix
match point set
