【摘要】 针对海量空间科学数据的精细及实时三维绘制需求,提出并实现了一种基于CUDA语言的并行化光线投射体绘制加速算法,利用传统体绘制算法中光线投射法的可并行特点和GPU中高速的纹理查询的优点,通过一个实际坐标到纹理坐标的转换函数实现了对不规则采样数据的准确采样,并完成了绘制算法的CUDA并行化改造,通过CUDA语言利用GPU强大的并行计算能力实现了对海量空间数据的实时三维光线投射绘制。更多还原

【Abstract】 Massive space science data needs fine and real-time three-dimensional rendering.In order to meet this requirement,a CUDA-based parallel accelerated ray casting algorithm is proposed and implemented.Since the traditional ray casting algorithm can be parallel and the GPU is capable of high speed texture mapping,this accelerated ray casting algorithm contains a parallel CUDA rendering algorithm which is modified from the traditional one and it also uses a real world to texture coordinates conversion function to achieve an accurate sampling on irregular sample data.The use of CUDA and GPU’s powerful ability on parallel computing makes real-time three-dimensional ray casting rendering of massive spatial data possible.更多还原