Moving with the flow: an automatic tour of unsteady flow fields

Abstract

We present a novel framework that creates an automatic tour of unsteady flow fields for exploring internal flow features. Our solution first identifies critical flow regions for time steps and computes their temporal correspondence. We then extract skeletons from critical regions for feature orientation and pathline placement. The key part of our algorithm determines which critical region to focus on at each time step and derives the region traversal order over time using a combination of energy minimization and dynamic programming strategies. After that, we create candidate viewpoints based on the construction of a simplified mesh enclosing each focal region and select the best viewpoints using a viewpoint quality measure. Finally, we design a spatiotemporal tour that efficiently traverses focal regions over time. We demonstrate our algorithm with several unsteady flow data sets and perform a user study and an expert evaluation to confirm the benefits of including internal viewpoints in the design.