一种无接缝纹理映射算法的并行实现

随着多核处理器的问世,并行编程变得简单和必要。通过对一种纹理映射算法的研究,对其耗时部分进行并行化,以加快算法运行速度。该纹理映射算法先建立三维网格顶点和视点之间的对应关系,然后利用重心坐标权重插值的颜色填充方法构造出一个无接缝的纹理图像。通过利用算法过程中存在的并行性,对算法进行并行实现。实验结果显示,能够得到一个无接缝的纹理图像,而且在双核处理器上的平均加速比达到了1.70。

Parallel Implementation of Seamless Texture Mapping Algorithm

With the advent of the multi-core processor, The writing of parallel programs becomes fairly easy and necessary. Based on the study of a texture mapping algorithm, its time-consuming part is paralleled, thus to hasten its operation speed. For this texture mapping algorithm, the corresponding relation between mesh vertex and the viewpoint is constructed first, then a seamless texture map implemented via barycentric coordinate weight color interpolation. By using the parallelism existing in the process of the algorithm, the algorithm is paralleled. The experimental results indicates that this method could achieve a seamless texture, and the average speed is up to 1.70 on a dual-core processor.