基于分布式渲染架构的远程可视化研究

互联网带宽的增长催生了远程可视化，它有着更好的分享性、移动性和方便性.针对大规模数据的远程可视化问题，提出了一种基于Sort-Last的分布式渲染架构，给出了基于GPU的融合、抗锯齿等算法.该架构用于远程可视化的服务器端，包括渲染节点、融合节点和任务节点等3层结构，具有良好的可扩展性.基于此，实现了一个远程可视化系统Waterman，提供基于Internet的高精地形渲染和海洋排放口污水扩散可视化服务，并给出了详细的设计方法和技术细节，包括基于Raycasting的地形渲染算法、基于陆地掩蔽(mask)方法的海面渲染技术和基于图片、网格模型的客户端混合实现技术等.最后对该架构和系统进行了性能测试和分析.提出的方法实用、鲁棒、扩展性好，可为同类系统设计提供很好的参考.

Remote Visualization Based on Distributed Rendering Framework

The progressive increase of Internet bandwidth makes remote visualization possible, which has advantages of remote share, mobility and convenience. A Sort-Last based distributed rendering architecture composed of rendering nodes, blending nodes and task nodes, and the corresponding algorithms including GPU based blending and anti-aliasing, are proposed to address the large-scale data rendering problem. The framework is highly parallel and extensible at each level such as rendering, blending and task balance, that is, arbitrary nodes can be deployed as stack structure according to performance requirement. On that basis, a remote visualization system, namely Waterman, is implemented to provide Internet services of terrain rendering and sewage visualization of sea outfalls. The design methods and techniques of Waterman are introduced in detail, which includes GPU based Raycasting algorithm for fast terrain rendering, land mask based technique to remove water area of aero photo texture for sea surface rendering and visualization, and hybrid rendering method using both polygon model and image based rendering. Finally, extensive experiments are settled to analyze benchmark of the distributed rendering framework, as well as time performance and data traffic throughput of the total remote visualization system. The results indicate that the proposed method is practicable, robust, and can serve as a good study case for similar systems.