基于矢量地图数据的异构混合快速绘制框架

A Heterogeneous Mixed Fast Drawing Framework Based on Vec-map Data

随着测绘、遥感及相关技术的发展,矢量地图的数据规模和细节层次迅速增加,它们含有大量复杂的地理空间要素,且要素间存在复杂的拓扑几何关系,这使得大规模矢量地图的快速可视化对于高性能并行计算的需求十分迫切。针对这一问题,本文在原有的CPU多核处理的基础上引入GPU加速运算,提出一种基于CPU-GPU协同工作的异构混合并行绘制技术。实验结果表明,这种处理框架在矢量数据的快速可视化中作用较为显著,在大比例尺数据下并行加速比达到了较高的数值。

With the development of mapping,remote sensing and related technology,the data scale and detail level of vector map are increasing rapidly. They contain a lot of complicated geospatial elements,and there are complex topological geometries among the elements. This makes high-performance parallel computing is very urgent for the rapid visualization of large-scale vector maps. Aiming at this problem,this paper introduces a GPU-accelerated operation based on the original CPU multi-core processing,and proposes a heterogeneous hybrid parallel rendering technique based on CPU-GPU cooperative work. The experimental results show that this kind of processing frame plays a more important role in the rapid visualization of vector data,and the parallel acceleration ratio achieves higher values under large scale data.