基于特征点的ICP点云配准算法  

ICP Point Cloud Registration Algorithm Based on Feature Points

作　　者：单丽杰 岳建平[1] 钱炜 SHAN Lijie;YUE Jianping;QIAN Wei(College of Earth Science and Engineering,Hohai University,Nanjing 211100,China)

机构地区：[1]河海大学地球科学与工程学院,江苏南京211100

出　　处：《甘肃科学学报》2022年第4期1-4,19,共5页Journal of Gansu Sciences

基　　金：国家重点研发计划(2018YFC1508603)。

摘　　要：ICP配准算法对待配准点云初始位置要求较高,且配准过程耗时长,因此提出了一种基于特征点改进的ICP点云配准算法。利用点云的局部法向量提取特征点,根据特征点的特征直方图得到初始配准点云;通过K-D tree搜索点集中的对应点对,运用四元数法得到配准参数,根据刚性距离约束条件精确配准点云。实验表明,该算法避免了ICP配准算法因初始位置姿态而陷入局部最优,同时提高了配准效率和精度。The ICP registration algorithm has higher requirements for the initial position and the registration process takes a long time.An improved ICP point cloud registration algorithm based on feature points is proposed in this paper.The local normal vector of the point cloud is used to extract the feature points,and the initial registration point cloud according to the feature histogram of the feature points is obtained.Then the corresponding point pairs in the point set are searched through the K-D tree,and the quaternion method is used to obtain the registration parameters.Finally according to the rigid distance constraint condition,the registration point cloud is figured out.Experiments show that the algorithm avoids the ICP registration algorithm from falling into the local optimum due to the initial position and attitude,and at the same time it improves the registration efficiency and accuracy.

关 键 词：点云配准 特征点 最近点迭代 K-D tree 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]