Abstract: 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.

Key words: vector data, rapid visualization, parallel computing, heterogeneous, CPU-GPU cooperation