针对传统的点特征点云配准算法不能解决缺失同名点的配准问题，从点在平面上的几何关系出发，利用具有较强几何约束能力的直线特征，推导出基于单位四元数表征的无需同名直线端点完全对应的点云配准模型。结果表明，该模型不仅能够同步求解配准参数，还适用于大旋转角的配准问题，同时降低了配准模型对参数初值的要求，增强了配准模型的稳定性和实用性。"/>
A Linear Feature Point Cloud Registration Algorithm Based on the Relation between Point and Plane" />
Aiming at the problem that the traditional point cloud registration algorithm cannot solve the registration problem of missing points with the same point, in this paper we derive a point cloud registration model using the linear features with strong geometric constraints. Our model is based on the geometric relationship of points on a plane, which is represented by unit quaternion and does not need to exactly correspond to the endpoint of a line with the same name. The experimental results show that this model can not only solve the registration parameters synchronously, but also can be applied to the registration problem of large rotation angles, which reduces the requirements of the registration model on the initial value of parameters and enhances the stability and practicability of the registration model.
Abstract：Aiming at the problem that the traditional point cloud registration algorithm cannot solve the registration problem of missing points with the same point, in this paper we derive a point cloud registration model using the linear features with strong geometric constraints. Our model is based on the geometric relationship of points on a plane, which is represented by unit quaternion and does not need to exactly correspond to the endpoint of a line with the same name. The experimental results show that this model can not only solve the registration parameters synchronously, but also can be applied to the registration problem of large rotation angles, which reduces the requirements of the registration model on the initial value of parameters and enhances the stability and practicability of the registration model.
