Skip to Main Content
Visualizing 2-D vector fields using streamlines is one popular flow visualization technique. Standard streamline generation algorithms compute the density of streamlines across the domain, detect features, and employ customized rules to emphasize features. In this process, feature characterization and visual clarity are heavily considered. Simultaneously preserving the temporal coherence for time-varying vector fields, however, remains a challenge. In this paper, we present a coherent and feature-aware streamline generation algorithm by employing a feature-guided streamline seeding technique and a coherent streamline placing scheme. For each frame, a feature map is first computed with critical points or the Finite-Time Lyapunov Exponent (FTLE) approach, and is used to initialize a set of seeds by leveraging the Poisson Disk distribution. These seeds are further optimized by using a deformation-driven moving mesh method. To preserve the temporal coherence, the streamlines generated from the seeds are individually checked subject to their correspondences to the ones in the previous frame. Subsequently, additional streamlines are sequentially inserted in low-density regions. We demonstrate our algorithm on both Computation Fluid Dynamics (CFD) and non-CFD datasets, and compare it with the recent literature.