Flow Feature Visualization Using Logical Operators on Multivariate Fields

Due to the large size of high-dimensional datasets that result from contemporary computational flow simulations, the classification and visualization of features is an essential, though challenging task for proper scientific analysis. We present a visualization system based on first-order fuzzy logic, that allows to convert natural language statements involving multiple scalar properties into well defined features which can be used for that allows to define and combine multiple feature criteria as logic visualization using geometric primitives in conjunction with the underlying flow field. A feature criterion can be defined as an atomic point predicate, which can be understood as a function, that maps all data points of a dataset to a Boolean value. Boolean algebra can then be used to combine these atomic predicates to define more complex ones. The combination of several feature criteria to one characteristic subset can be used to build one single geometric isosurface representation of several features, and thus, significantly reduce the amount of graphical primitives needed to display all features separately, minimizing clutter and occlusion. Further, the created subset can be utilized for particle seeding, with the aim to show the behavior of the flow in the surrounding area. We evaluate the positive and negative aspects of two different types of logical operators for the example of different simulation datasets and several feature criteria.