《计算机应用研究》|Application Research of Computers

基于GPU并行的点云数据简化的改进算法

Improved algorithm for point cloud data simplification based on GPU parallel

免费全文下载 (已被下载 次)  
获取PDF全文
作者 李普山,李伟波,冯智莉,万权,王海荣
机构 武汉工程大学 a.智能机器人湖北省重点实验室;b.计算机科学与工程学院,武汉 430205
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2020)09-033-2730-04
DOI 10.19734/j.issn.1001-3695.2019.04.0111
摘要 受环境因素影响,卤水下矿床表面地势平缓,采集的矿床点云冗余点较多,为了提高对矿床进行三维建模的效率,设计了一种基于GPU并行的点云简化的改进算法。对每个小栅格内的点进行最小二乘的平面拟合,根据各个点到拟合平面的距离精简了大部分冗余点,并通过剩余点的曲率进行了第二次精简。将整个处理过程限定在每个小栅格内,在降低计算量的同时避免了因过度简化而出现的空洞现象。另外,对点云的简化过程进行了基于GPU的多线程并行处理,极大地提高了整个处理过程的效率。实验表明,算法改进后达到原算法效果的同时提高了算法效率,利用GPU加速后,大大缩短了算法的执行时间。
关键词 点云简化; 栅格法; 平面拟合; GPU并行
基金项目 湖北省高校产学研合作重点资助项目(C2010033)
本文URL http://www.arocmag.com/article/01-2020-09-033.html
英文标题 Improved algorithm for point cloud data simplification based on GPU parallel
作者英文名 Li Pushan, Li Weibo, Feng Zhili, Wan Quan, Wang Hairong
机构英文名 a.Hubei Key Laboratory of Intelligent Robot,b.School of Computer Science & Engineering,Wuhan Institute of Technology,Wuhan 430205,China
英文摘要 The surface of the submerged ore deposits is flat and the collected points have more redundant points. In order to improve the efficiency of 3D modeling for deposits, this paper designed an improved algorithm for point cloud simplification based on GPU parallel. It fitted the points in each small grid by the least squares method, simplified the most of the redundant points according to the distance from each point to the fitting plane, and simplified the curvature of the remaining points at the second time. It confined the whole process to each small grid, which reduced the computational complexity and avoided the cavity phenomenon caused by over simplification. In addition, it processed the simplified process of point cloud in parallel with multi-threads based on GPU, which greatly improved the efficiency of the whole process. Experiments show that the improved algorithm achieves the effect of the original algorithm, improves the efficiency of the algorithm, and greatly reduces the execution time of the algorithm after using GPU acceleration.
英文关键词 point cloud simplification; grid method; plane fitting; GPU parallel
参考文献 查看稿件参考文献
 
收稿日期 2019/4/18
修回日期 2019/6/4
页码 2730-2733
中图分类号 TP301.6
文献标志码 A