Adaptive Distance Grid Based Algorithm for Farthest Point Seeding Streamline Placement

Abstract

Streamlines are commonly used in scientific visualization. They are the most used geometric items in primitive-based visualization algorithms. In this paper, a modified version of the farthest point seeding strategy streamline placement is presented. The main advantage compared to the original method is the use of a simple easy to implement underlying data structure. The Delaunay triangulation used in the original algorithm is heavy and susceptible to the calculation robustness errors. The distances between lines and the localization of the biggest void in the domain are approximated using the Delaunay triangulation. This paper presents an adaptive distance grid to model the visualization domain and incorporate anywhere the local distance to all the other streamlines and the boundaries exactly without any approximation. The streamlines are started at the farthest point from all the existing ones and the domain boundaries. The localization of the seed point position is directly accessible via the distance adaptive grid. The grid update is direct too, and is done by a greedy algorithm avoiding any additional cost. The obtained results are sufficient, and the extension to multi-resolution is straight and simple.

Keywords

• 2d streamline placement
• farthest point seeding strategy
• adaptive distance grid