Image-based flow visualization
 | This article's tone or style may not reflect the encyclopedic tone used on Wikipedia. |
| This article uses first-person (I, we) or second-person (you) inappropriately. (September 2011) |
In scientific visualization, image-based flow visualization is a technique proposed by Jarke van Wijk [1] to visualize flows, like the wind movement of a tornado. Compared with classical Integration like techniques it has the advantage of producing a whole image at every step. It is a method from the texture advection family.
Principle
The main idea is to create a noise texture on a regular grid and then bend this grid according to the flow (the vector field). We then sample the bended grid at the original grid location. The output is then an according to the flow displaced version of the noise. To use this for visualization we can for example do this for a few steps and make an animation or also inject some dye on a regular grid and see how this dye got transported away by the flow.
The advantage of this approach is that it can be accelerated on modern graphics hardware, thus allowing for real-time or almost real-time simulation of 2D flow data.
References
- ^ van Wijk, Jack (2002). "Image Based Flow Visualization" (PDF). Proceedings ACM SIGGRAPH 2002, San Antonio, Texas.