Point-Based Rendering of Forest LiDAR

Airborne Light Detection And Ranging (LiDAR) is an increasingly important modality for remote sensing of forests. Unfortunately, the lack of smooth surfaces complicates visualization of LiDAR data and of the results of fundamental analysis tasks that interest environmental scientists. In this paper, we use multi-pass point-cloud rendering to produce shadows, approximate occlusion, and a non-photorealistic silhouette effect which enhances the perception of the three-dimensional structure. We employ these techniques to provide visualizations for evaluating two analysis techniques, tree segmentation and forest structure clustering.