基于点云数据的重建算法研究.pdf

摘 要
近年来随着快速 、高精度的三维数字化扫描设备的不断涌现, 人们获取物体表面采
样点数据变得相对简单 、 便宜而且准确 。将这些采样数据集转换成几何模型的过 程称为
曲面重建, 该技术可广泛应用于计算机辅助几何设计、 计算机图形学 、 医学图像处理 、
计算机视觉等领域。本文从三维数字化扫描设备获取的数据出发, 对点云数据重建方法
进行了研究, 具体工作如下 :
针对点云的曲面重建, 本文分析了国内外点云的重建算法, 比较了它们的优点和不
足, 汲取这些算法的优点, 提出了一种边界保持的点云数据重建算法 。该算法根据拟合
误差对点云进行自适应拟合, 通过减少对八叉树高层节点的拟合次数, 提高了算法效率 ;
并修正单元分解方法中的权函数, 加入密度权函数, 削弱点云密度的不一致对重建曲面
影响。在生成三角网格时, 结合原始的点云数据, 对所得网格进行边界约束 。实验结果
表明, 该算法原理简单, 重建速度快, 重建效果良好, 满足三维人脸整形与美容虚 拟手
术系统对曲面重建的要求。
关键词 点云, 曲面重建, 三角化, 边界保持
I
A BSTR A C T
In recent years, w ith the fast, high precision of 3d digitizing scanning equipm ent
constantly emerged; people obtain surface sampling data become relatively sim ple, cheap and
accurate. The process that convert sam pling data to geometric m odel called surface
reconstruction, the technology can be widely applied to com puter aided geometric design,
computer graphics, m edical image processing, computer vision, etc. T his paper researched
cloud data reconstruction m ethods based on data acquired from the 3d digitizing scanning
equipment, and the specific w ork are as folow s:
A iming at the surface reconstruction of point cloud, this paper analyzes the domestic and
foreign point cloud reconstruction algorithm, and compares their advantages and
disadvantages, absorbing the advantages of these algorithm s, and puts forw ard a kind of
boundary maintenance point cloud data reconstruction algorithm.
According to the fitting error on the point cloud, this algorithm w ill be self-adapti