A new line integral convolution algorithm for visualizingtime-varying flow fields

摘要

New challenges on vector field visualization emerge as timendependent numerical simulations become ubiquitous in the field ofncomputational fluid dynamics (CFD). To visualize data generated fromnthese simulations, traditional techniques, such as displaying particlentraces, can only reveal flow phenomena in preselected local regions andnthus, are unable to track the evolution of global flow features overntime. The paper presents an algorithm, called UFLIC (Unsteady Flow LIC),nto visualize vector data in unsteady flow fields. Our algorithm extendsna texture synthesis technique, called Line Integral Convolution (LIC),nby devising a new convolution algorithm that uses a time-accurate valuenscattering scheme to model the texture advection. In addition, ournalgorithm maintains the coherence of the flow animation by successivelynupdating the convolution results over time. Furthermore, we propose anparallel UFLIC algorithm that can achieve high load balancing fornmultiprocessor computers with shared memory architecture. We demonstratenthe effectiveness of our new algorithm by presenting image snapshotsnfrom several CFD case studies