2D Unsteady Flow Visualization by Animating Evenly-Spaced Streamlets
Flow visualization has been widely used to display and discover patterns and features in vector fields. Common applications include the representation of ocean currents and weather model data. In this thesis, a flexible method for animating vector fields is developed, based on a generalization of a Poisson disc sampling method. An image rendering technique is generalized so that it can be applied to the generation of shaped streamlets with minimal overlap. The algorithm has two stages, in the first streamlets are drawn into an image buffer, larger than their intended size. Before they are drawn they are tested to see if they impact on already drawn areas and if they do they are rejected. In the second stage the ones that pass the test are drawn normal size. The concept of 3D streamlet object, which groups couple consecutive time step streamlets as a primitive rendering object, is introduced as part of a method for animating streamlets so that they have minimal overlap and show frame-to-frame coherence providing visual continuity when animating time varying vector fields. Acceptance schemes that allow for occasional overlap between streamlets are explored and found to improve both the speed and the overall quality. Both model data and real weather data are used to evaluate the method. The results show that the method is flexible, allowing for variable size and density of streamlets, and produced good results.