一种基于建模的图像渲染算法的并行实现

Parallel Implementation of Image Rendering Algorithm based on Modeling

提出了一种基于建模的图像渲染技术的并行实现方法,在已知目标3D模型和适量环绕视点图像的前提下,通过三角剖分法生成虚拟视点,并插值生成虚拟视点下的采样信息,对3D模型的表面进行渲染,最终达到自由视点下漫游目标的目的。为加快程序运行,采用OpenMP对算法的关键部分进行了并行化优化。实验表明,串行算法可以获得比较满意的渲染结果,并且在不影响实验结果的前提下,算法的并行部分在双核处理器上的平均加速比达到了1.70。

An algorithm based on the modeling of image rendering technology is proposed In the known objective＇ s 3D model and moderate surround view images, the virtual view and its sample information are generated for rendering the 3D model, and finally the purpose of target roaming under arbitrary viewpoint is achieved. Meanwhile, OpenMP is adopted to deserialize key part of the algorithm and thus to speed up the program operation. The experiment shows that this algorithm could achieve satisfactory rendering effects, and the average acceleration ratio of parallelization algorithm could reach 1.70 in dual-core processor.