基于多分辨方法的海量数据三维云体绘制

Three-dimensional Cloud Rendering of Massive Data Based on Multi-resolution Method

针对因图形硬件存储能力有限带来的海量云数据体绘制的交互性差等问题,提出了一种适用于大规模体数据压缩的多分辨率三维云体绘制方法。该方法首先引入基于变异系数的多分辨率细节层次划分,再利用小波变换和最优量化对划分数据块进行压缩,最后采用基于GPU的自适应采样光线投射算法对云数据进行体绘制。实验结果表明:该方法与传统的多分辨率压缩绘制方法相比,既提高了绘制速度,又保证了绘制的逼真效果。

Aiming at the problem of poor interactivity of massive cloud data volume rendering due to limited storage capacity of graphics hardware,a multi-resolution volume rendering method for large-scale volume data compression was proposed.Firstly,the multi-resolution detail hierarchy based on coefficients of variation was introduced.The wavelet transform and optimal quantization were used to compress the partitioned data blocks.Finally,the GPU-based adaptive sampling ray casting algorithm was used for cloud data volume rendering.The experimental results show that compared with the traditional multi-resolution compression rendering methods,the proposed method can not only improve the rendering speed,but also ensure the realistic effect of drawing.