一种高效的支持向量回归三维点云修补算法  

Efficient point cloud reconstruction algorithm based on support vector regression

作　　者：杨蕾[1] 

机构地区：[1]西南科技大学制造科学与工程学院,四川绵阳621010

出　　处：《计算机应用研究》2009年第10期3945-3947,3959,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(10576027)

摘　　要：给出了一种基于支持向量回归的三维点云空洞修补算法,该算法首先将残缺区域边界点集向邻近区域的切平面投影,投影点集作为训练数据集,通过支持向量回归,得到残缺区域所服从的隐式曲面方程,完成修补。为提高算法效率,将该修补问题转换为等价的最小包含球问题,降低了算法的复杂度。该修补算法能够较好地使修补点云与原始点云平滑融合,具有很好的恢复效果。This paper proposed an efficient point cloud reconstruction algorithm based on support vector regression. Firstly, projected the boundary points of the hole on the tangent plane, obtained the implicit surface equation of the hole by support vector regression using the training points projected on the tangent plane. Then, got the restoration points of the hole through the surface equation. In order to improve the efficiency of the algorithm, converted the restoration problem to minimum enclosing ball problem. The proposed reconstruction algorithm can restore the hole of the point cloud smoothly, and has good restoration results.

关 键 词：反求工程 点云 支持向量回归 最小包含球问题 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]