基于双缓冲队列的海量地形数据并行处理方法-《郑州大学学报(工学版)》

文章信息/Info

Title:: Double-Buffered Queue-Based Parallel Processing Algorithmfor Massive Terrain Dataset

作者:: 渠涧涛; 陈小潘; 赵亚萌; 王鹏; 陈玉林; 1.河南大学计算机与信息工程学院,河南开封475004;2.中国科学院遥感与数字地球研究所,北京100012

Author(s):: CHEN Xiaopan1 ; QU Jiantao1; 2; ZHAO Yameng2; WANG Peng1; 2 ; CHEN Yulin1; 1. School of Computer and Information Engineering, Henan University, Kaifeng 475004, China;2. Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing 100012, China

Keywords:: massive terrain data; double buffer queue; parallelization; large-scale terrain rendering; tile load

摘要:: 海量地形数据处理过程中,硬件的性能优势不能得到充分利用,已成为制约大规模地形绘制速度的瓶颈.针对这一问题,分析了影响硬件处理能力发挥的关键因素,采用双缓冲队列的设计思路,提出了一种支持海量地形数据的并行处理方法,即将大规模地形绘制分为数据处理和渲染绘制两个独立的过程,分别进行并行处理:设立两个缓冲队列,将数据的读、写操作区分开来;充分考虑了瓦片加载的优先级,并据此进行任务分配.实验表明,该方法有效地提升了大规模地形绘制的整体速度.

Abstract:: When dealing with massive terrain data ，the advantage of hardware performance can’t be fully uti-lized. This has become a bottleneck，which restricts the speed of massive terrain tiles rendering. This paperanalyzes the key factors that affect large-scale terrain rendering speed，and proposes a parallel algorithm formassive terrain data processing. The algorithm adopts double buffer queues and divides large scale terrain ren-dering into two parallel processing which includes data processing and rendering. The two buffer queues areresponsible for data reading and writing operations in turn. The loading priority of terrain tiles is consideredand tasks are allocated based on the priority. The experimental results show that this approach improves thespeed of rendering massive terrain tiles greatly.

[1]应申,靳凤攒,李霖,等．基于ArcGIS Engine的矢量数据分层分块技术研究[J]．测绘地理信息,2014,39(6) :50 - 53.

[2]胡鹏昱，王晓军．基于Geometry Clipmap算法的大规模地形可视化研究[J]．计算机应用与软件，2012 ,29(7) :87 - 90.

[3]张俊峰,姚志宏．基于四叉树孤立分割和屏幕误差的地形LOD 算法[J]．西南交通大学学报,2013,48( 4 ) :666 -671.