基于集群机并行绘制的三角形条带压缩

Triangle Strip Compression based on Cluster Parallel Rendering

提出一个适用于集群机并行绘制的三角形条带数据压缩框架——视点连贯性的分片条带压缩(VCPSC),有效地克服了传统几何数据压缩算法存在的问题.VCPSC包括3步核心算法:基于空间和法向连贯性分片方法;基于同心圆全局路径控制的三角形单条带化;ETSC三角形条带压缩算法.通过把每个压缩的三角形条带映射为一个支持随机存取的虚拟三角形,VCPSC实现了几何模型压缩域的基于视点的归属判断和分片随机存取.实验结果表明:VCPSC有效地改善了集群机绘制性能.

We presented a triangle strip compression framework called visibility-coherently piecewise strip compression （VCPSC）, which eliminated the traditional inherent deficiency and suited for cluster rendering system. It included three steps： visibility and space coherence based segmentation method, concentric circle based single triangle strip generation and a single triangle strip compressed method ETSC （efficient triangle strip compress）. Since an encoded patch can act as a virtual primitive which can be randomly accessed, the encoded mesh can be partitioned in the encoded domain. As a sequential, every encoded patches of VCPSC can be view-dependent sorted without being decoded first. Experimental results show that VCPSC can dramatically improve the rendering rate.