Research And Implementation Of Parallel Volume Rendering Techniques And Visualization System For 3D Data Fields

Data field visualization technique is one of the most active research topics in computer graphics. Scientific visualization provides an intuitive and effective method to explore the information contained in 3D complex data. It can be applied in wide area, such as medical image processing, meteorology and computational fluid dynamics. This dissertation firstly introduces the basic concepts and methods of scientific visualization and parallel rendering techniques, surveys the new research development of parallel volume rendering techniques, and concentrates its research on parallel volume rendering techniques and visualization systems. A series of algorithms have been proposed, and several visualization systems have been designed and implemented.1. The dissertation proposes a parallel shear-warp volume rendering algorithm based on dynamic data distribution. Making use of the idle broadcast communication line, the algorithm performs data distributing and rendering in parallel. While keeping the previous algorithm's efficiency, the algorithm avoids redundant data storage. The improved task distribution strategy avoids imbalance load between computing nodes and job jams in pipeline.2. The dissertation implements a distributed sheared volume buffer(DSVB) based on cache-coherent sheared memory model. A parallel hybrid beam ray-tracing algorithm, which makes use of the ray coherence during the tracing process, is proposed to improve cache efficiency. A new dynamic load balancing method based on Recently-Loaded List is given to improve processor utilization. The spatial data structure based on data pages and optimized intersection method between rays and volume are proposed to reduce remote data access.