GPU-accelerated 2D point cloud visualization using smooth splines for visual analytics applications

Abstract

We develop an efficient point cloud visualization framework. For efficient navigation in the visualization, we introduce a spline-based technique for the smooth approximation of discrete distance field data. Implemented on the GPU, the approximation technique allows for efficient visualizations and smooth zooming in and out of the distance field data. Combined with a template set of prede-fined, automatically or interactively adjustable transfer functions, the smooth distance field representation allows for an effective visualization of point cloud data at random abstraction levels. Using the presented technique, sets of point clouds can be effectively analyzed for intra- and inter-point cloud distribution characteristics. The effectiveness and usefulness of our approach is demonstrated by application on various point cloud visualization problems

Feature-based automatic identification of interesting data segments in group movement data

“…Point groups can be represented by solid shapes such as hulls, 15,16 by various geometric constructs such as isocontours, 17 or using distance fields. 18 The time dimension of the data is usually shown by animation, 19 trajectories, 20 or a combination of both. 21,22 An empirical comparison of animation and trajectory techniques was presented in the study by Robertson et al, 23 concluding that for presentation, animation is more effective, and for detailed analysis, trajectories are more effective.…”

Visual analytics of time dependent 2D point clouds

“…In [SSZW08], an algorithm for construction of compact, enclosing shapes was presented. Distance fields allow representation of point sets by smooth formation of visual areas by using appropriate transfer functions [KTSZ08]. On the efficiency side, the visualization of massive point cloud data sets may be accelerated by appropriate data structures as presented in [HE03].…”