Despite the vast processing power of modern GPUs, rendering terrain geometry at sufficient frame rates is still a very difficult task. This is mainly due to the large terrain data that must be managed, displayed and modified at high frame rates. While traditional level-of-detail algorithms for terrain utilize most of the CPU and a fraction of the GPU, our algorithm exploits the parallel processing capabilities of the GPU. We represent the terrain geometry as a set of geometry images which allows us to apply image processing techniques to preserve terrain features during level-of-detail transitions and to render terrain that has realistic geographical features such as overhangs or caves. Shader technology is used to alter the terrain mesh at runtime based on the geometry images to create a continuous level-of-detail. Our contribution is a continuous level-of-detail algorithm that is accelerated by the GPU and focuses on preserving features of the terrain.
展开▼
