摘 要：本文提出了义齿海量点云数据的三维重建算法，包括隐函数的参数选择和局部曲面重建方法。在参数选择中从几种基函数的比较中确定三调和样条函数适合义齿重建要求。根据义齿的特征将义齿点云数据划分到若干个子空间，并为子空间对角线的交点附件的原始点增加偏移点，与传统方法相比较计算量降低了。该方法更适用于海量点云数据的三维重建，仿真实验结果表明本文方法在义齿三维重建上是有效的。
关键词：义齿模型；三维重建；隐函数；子空间
中图分类号：TP391.72 文献标识码：A 文章编号：2096-4706（2018）05-0186-03
Application Research of Three-dimensional Reconstruction in Denture Model
LIU Yanju1，LI Baiquan2，REN Hang1，LIU Yanzhong1，LIU Xiangjuan1
（1.College of Computer and Control，Qiqihar University，Qiqihar 161006，China；2.The Third Affiliated Hospital of Qiqihar Medical College，Qiqihar 161006，China）
Abstract：A surface reconstruction algorithm for mass point clouds of denture is proposed in this paper，which involves parameters selection of implicit function and local surface reconstruction. For parameters selection，three hamonic spline function is determined as basis function of implicit after comparing several functions and offset points are selected in unit space by the center intersection of main diagonal. According to the features of the denture，the data of the denture point cloud is divided into several subspaces，and the offset point is added to the original point of the intersection point of the diagonal line of the subspace，and the calculation is reduced compared with the traditional method. The method is suitable to reconstruct surface for mass point clouds. The experimental results demonstrate that the method is effective in denture surface reconstruction.
Keywords：denture model；three-dimensional reconstruction；implicit function；sub-space
参考文献：
[1] 刘艳菊，张永德，姜金刚. 三维点云模糊分类的法向量估值算法 [J]. 华中科技大学学报（自然科学版），2013，41（8）：50-54．
[2] Liu Yanjv，Miao Fengjuan et al. A novel self-organizing fuzzy neural network surface reconstruction for mass point clouds of irregular mode [J].ICIC Express Letters，2015，6（9）：2377-2383.
[3] 杨剑，宋超峰，宋文爱，等. 基于遗传算法的模糊RBF神经网络对遥感图像分类 [J]. 小型微型计算机系统，2018，39（3）：621-624.
[4] 张奎，王建南，王肖峰. 基于神经网络的变压器故障诊断 [J]. 电子测量技术，2017，40（12）：98-101．
作者简介：刘艳菊（1974-），女，黑龙江齐齐哈尔人，教授，博士，研究方向：三维重建、智能控制及口腔医学图像。