基于视点运动的海量数据快速显示方法研究

Out-of-core Data Fast Display Method Based on the Viewpoint Movement Research

为了海量数据实时准确的显示,提出一种基于视点运动模型的可视化算法。该算法首先采用Hilbert-R树对海量数据建立快速索引,然后根据视点运动模型将数据块分为可视域、潜在域、卸载域。最后按照索引将相应数据块导入内存,可视域数据块会被进一步载入GPU进行绘制。当视点位置发生变化时,GPU直接加载位于内存的潜在域数据块。实验结果表明,在保证渲染质量的前提下,该算法能够快速显示海量数据体并实现实时交互,显示速度提升36%~78%。

In order to accurately display the real-time out-of-core data,a visualization algorithm based onthe viewpoint of motion model is presented.Firstly,Hilbert-R tree is used to index data.Then,the datablock can be divided into the visible area、the potential area、the unloading area according to theviewpoint motion model.Finally,the corresponding data is imported into memory according to the index.Data blocks in the visible area will be further loading of GPU rendering.When the viewpoint changes,thepotential data block in memory is loaded directly into the GPU.The experimental results show that underthe premise of ensuring the quality of rendering,out-of-core data can be fast displayed and interactive,the display speed was raised36%~78%.