Multi-Domain Correlation for Vortex Extraction in Fluid Flow Fields

Effective extraction of characteristic structures of flow features, such as vortices, can simplify visualization and thus potentially enable users easier cognition of flow field data. In this paper we present a novel method to extract vortex structures through correlation of flow domains transformed from the velocity field. This correlation is based upon our earlier "maxima score" approach, which is domain-independent and normalized. By correlating the maxima score of flow domains, we have improved its capabilities for extracting characteristic structures of vortices. We have compared our method with the feedback of flow analysis experts, and verified the capabilities of our method using flow data derived from computational fluid dynamics and experimental measurements. These results show that this new method reduces noise, and operates at a higher sensitivity for vortex extraction than the maxima score is capable of alone.