颅骨点云模型的优化配准

发布时间：2018-10-26 06:53

【摘要】：由于颅骨的三维点云数据模型复杂且不同人的颅骨差异较小,对其配准精度要求较高。为了提高颅骨点云模型的配准精度和收敛速度,提出了一种先粗配准再细配准的配准方法。首先,对颅骨点云数据模型进行去噪、简化和归一化等预处理;然后,通过区域划分、区域配准和求解组合系数以及求解刚体变换等步骤实现区域层次上的颅骨粗配准;最后,通过引入动态迭代系数来改进基于旋转角约束的迭代最近点算法,并采用该改进的ICP算法实现颅骨的细配准,从而达到精确配准的目的。实验结果表明:与ICP算法相比,改进的ICP算法的配准精度和收敛速度分别提高了约30%和50%。证明该种先粗配准再细配准的颅骨点云模型配准方法是一种精度高、速度快的有效颅骨配准算法。
[Abstract]:Because the 3D point cloud data model of skull is complex and the difference of skull is small, the registration accuracy is high. In order to improve the registration accuracy and convergence rate of the skull point cloud model, a registration method of coarse registration and fine registration was proposed. Firstly, the cranial point cloud data model is pretreated with de-noising, simplification and normalization, and then the rough registration of skull at the regional level is realized through the steps of region division, region registration, solving the combination coefficient and solving the rigid body transformation. Finally, the iterative nearest point algorithm based on rotation angle constraint is improved by introducing dynamic iterative coefficients, and the improved ICP algorithm is used to realize the fine registration of the skull, so as to achieve the purpose of accurate registration. The experimental results show that the registration accuracy and convergence speed of the improved ICP algorithm are improved by about 30% and 50%, respectively, compared with the ICP algorithm. It is proved that the method of cranial point cloud registration with coarse registration and fine registration is an effective skull registration algorithm with high accuracy and high speed.
【作者单位】： 咸阳师范学院教育科学学院;西北大学信息科学与技术学院;北京师范大学信息科学与技术学院;
【基金】：国家自然科学基金资助项目(No.61373117,No.61305032)
【分类号】：TP391.7
，

本文编号：2294913