| 一种基于CUDA的三维点云快速光顺算法 |
| |
| 引用本文： | 唐杰,徐波,宫中樑,武港山. 一种基于CUDA的三维点云快速光顺算法[J]. 系统仿真学报, 2012, 24(8): 1633-1637,1642 |
| |
| 作者姓名： | 唐杰 徐波 宫中樑 武港山 |
| |
| 作者单位： | 南京大学软件新技术国家重点实验室,南京,210093 |
| |
| 基金项目： | 国家高技术研究发展计划(863)(2007AA06A402);国家科技重大专项(2011ZX05035-004-004HZ) |
| |
| 摘 要： | 提出了一种基于CUDA的点云光顺算法。算法细分成点云空间划分,K邻近搜索,法矢估算以及光顺等四个独立的且并行程度非常高的步骤。设计了基于CUDA的点云空间平均单元格划分算法及数据结构,有效提升了点云的划分效率;设计了基于CUDA的空间K邻近搜索算法;改进了点云法矢估算方法,提出了高斯加权的法矢计算方法,有效改善了法矢估算效果;在光顺过程中加入了邻近点的面积影响因子,缓和了过光顺等不足。最后通过实验验证了算法的有效性。
|
| 关 键 词： | 光顺 CUDA GPU计算 点云 |
Fast Fairing of 3D Point Clouds Using CUDA |
| |
| TANG Jie,XU Bo,GONG Zhong-liang,WU Gang-shan. Fast Fairing of 3D Point Clouds Using CUDA[J]. Journal of System Simulation, 2012, 24(8): 1633-1637,1642 |
| |
| Authors: | TANG Jie XU Bo GONG Zhong-liang WU Gang-shan |
| |
| Affiliation: | (National Key Laboratory for Novel Software Technology,Nanjing University,Nanjing 210093,China) |
| |
| Abstract: | A CUDA-based point cloud fairing algorithm was proposed.The algorithm is composed of four steps with great parallelism including point cloud space partitioning,K-nearest neighbors searching,the normal estimation and fairing.A CUDA-based point cloud partition method as well as its data structure which utilizes the uniform grid was designed,which improved the efficiency of partitioning greatly.A CUDA-based algorithm for K-nearest neighbors search was designed.An improved normal estimation method was proposed which utilized Gaussian weighted method to calculating normal vector and improved the precision of normal estimation.The impact factor of the adjacent area was introduced to improve the effect of smoothing and alleviate the degree of over smoothing.Finally,the experiments verify the effectiveness of the algorithm. |
| |
| Keywords: | fairing CUDA GPU computing point clouds |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
