SBIR Phase I: Visualization of Massive Multivariate Adaptive Mesh Refinement (AMR) Data

该小型企业创新研究(SBIR)第一阶段项目解决了通过自适应网格细化(AMR)过程创建的分层重叠结构网格缺乏可视化技术的问题。虽然AMR结构使得使用统一网格方法进行模拟成为可能，否则这些模拟的计算量和内存成本都太高，但它使科学家缺乏可视化工具来正确渲染产生的体数据。因此，需要专注于体积绘制的可视化工具，这些大型、多变量、时变的数据是由使用AMR技术的模拟产生的。在这个项目中，描述了一种体绘制方法，它将使科学家能够有效地可视化他们的模拟结果。将实施光线投射算法，以实现准确的可视化，并能够以质量换取速度、多变量数据的灵活合成以及从大规模并行到台式计算机的各种计算平台的适应性。将创建一个原型可视化系统，该系统将通过使用细节层次技术、对多变量数据的高级传递函数编辑和体绘制解决方案与标准几何技术的混合来演示交互性。在开发周期的早期，核心的AMR可视化技术将被整合到开源可视化工具包中，以获得有价值的反馈并在该领域获得认可。这将为研究人员提供早期访问这些可视化技术的机会，从而腾出当前用于实施定制解决方案的宝贵时间。这些AMR可视化方法作为源代码的可用性将使AMR可视化领域的未来研究能够基于通过这一努力获得的结果。将更新可视化工具包的文件(包括商业教科书和用户指南)和培训课程，以包括AMR可视化，为该领域的新研究人员提供教育机会。除了作为开放源码工具包的一部分提供这项技术外，还将开发一个商业应用程序，以在交钥匙应用程序中提供AMR可视化。在终端用户应用程序中提供的AMR可视化将消除大多数研究人员开发自己的可视化解决方案的需要，从而有更多的时间专注于核心科学。

This Small Business Innovation Research (SBIR) Phase I project addresses the lack of visualization technology for hierarchical overlapping structured grids created through a process known as Adaptive Mesh Refinement (AMR). Although the AMR structure makes possible simulations that would have otherwise been too computationally and memory expensive using a uniform grid approach, it leaves the scientist with a lack of visualization tools to properly render the resulting volumetric data. Therefore, there is a need for visualization tools that are focused on volume rendering the large, multivariate, time-varying data produced by simulations using the AMR technique. In this project, a volume rendering approach is described that will enable scientists to effectively visualize the results of their simulations. A ray casting algorithm will be implement allowing for accurate visualization with the ability to trade quality for speed, flexible compositing of multivariate data, and adaptability across a variety of computational platforms from massively parallel to a desktop computer. A prototype visualization system will be created which will demonstrate interactivity through the use of level of detail techniques, advanced transfer function editing for multivariate data, and intermixing of the volume rendering solution with standard geometric techniques.Early in the development cycle, the core AMR visualization technology will be incorporated into the open source Visualization Toolkit in order to gain valuable feedback and to gain recognition in this field. This inclusion will provide researchers with early access to these visualization techniques, freeing up valuable time currently spent implementing custom solutions. The availability of these AMR visualization methods as source code will enable future research in the area of AMR visualization to be based upon the results obtained through this effort. The documentation (including a commercially available textbook and users' guide) and the training classes for the Visualization Toolkit will be updated to include AMR visualization, providing educational opportunities for new researchers in the area. In addition to the availability of this technology as part of an open-source toolkit, a commercial application will be developed to provide AMR visualization in a turn-key application. AMR visualization provided in an end-user application will eliminate the need for most researchers to develop their own visualization solutions, allowing more time to be focused on the core science.