Efficient level of detail for texture-based flow visualization

详细信息    查看全文

• 作者：Daying Lu ; Dengming Zhu ; Zhaoqi Wang and Jinzhu Gao
• 关键词：level of detail ; texture-based visualization ; noise optimization ; efficient acceleration ; smooth animation
• 刊名：Computer Animation and Virtual Worlds
• 出版年：2016
• 出版时间：March/April 2016
• 年：2016
• 卷：27
• 期：2
• 页码：123-140
• 全文大小：1898K
• ISSN：1546-427X

文摘

In this paper, we present an efficient level of detail algorithm for texture-based flow visualization. Our goal is to enhance visual perception and performance and generate smooth animation. To achieve our goal, we first model an adaptive input texture taking into account flow patterns to output view-dependent high-quality images. Then, we compute field lines only from sparse sampling points of the input noise texture for outputting volume line integral convolution textures and skip empty space utilizing two quantized binary histograms. To improve image quality, we implement anti-aliasing through adjusting the line integral convolution step size and thickness of trajectory lines with an opacity function. We further extend our solution to unsteady flow. Flow structures and evolution are clearly shown through smooth animation achieved with coherent evolution of particles, handling of discontinuous flow lines, and spatio-temporal linear constraint of the underlying noise volume. In the result section, we show high-quality level of detail of three-dimensional texture-based flow visualization with high performance. We also demonstrate that our algorithm can achieve smooth evolution for unsteady flow with spatio-temporal coherence. Copyright