海量地形数据组织与调度策略研究  

Research on organization and scheduling strategy of massive terrain dataset

作　　者：袁松鹤 靳海亮[1] 耿文轩 YUAN Songhe;JIN Hailiang;GENG Wenxuan(Henan Polytechnic University,Jiaozuo 454000,China)

机构地区：[1]河南理工大学,河南焦作454000

出　　处：《测绘工程》2018年第8期31-36,共6页Engineering of Surveying and Mapping

基　　金：国家自然科学基金资助项目(41541011)

摘　　要：针对海量地形数据与有限的计算机内存之间的矛盾问题,需对数据的组织与调度过程进行优化。对海量的地形数据基于分块的方法以四叉树的数据结构进行组织。以块内无冗余的数据存储方式构建多分辨率层次模型,同时生成层次索引和坐标序号索引。在预处理阶段进行自上而下的参数计算,以此种方式可以快速读取到需求数据。其次根据视点位置与视线的方向两个因素,利用视锥体投影的方式对数据进行预取,实施两次数据裁剪策略,并结合多线程策略实现数据的加载与卸载。实验结果表明该策略具有较高的实时性,能够满足大规模地形数据的高效绘制要求。Concerning the contradiction of the huge terrain data the limited computer memory and the limited computer memory,the process of data organization and scheduling is optimized.Based on the method of block,this paper uses the data structure of quadtree to organize the massive terrain data.A model of multi-resolution based on non-redundant data storage method in block is built,at the same time,the index of level and coordinate number is also generates.In order to get the demand data quickly,it presented the calculation method of top-down parameters in the preprocessing stage in advance.And then a prefetching scheme of the view frustum projection is proposed according to eyes location and the direction of sight-line,which implements data prefetching by using two-data cutting.In combination with multi-thread,the data loading and unloading is accomplished.The experimental result shows that this method achieves good real-time performance which can meet the requirements massive terrain data visualization.

关 键 词：四叉树数据结构 块内无冗余 多分辨率模型 视锥投影 多线程 

分 类 号：P208[天文地球—地图制图学与地理信息工程]