Animation of trees in wind using sparse motion capture data

摘要

We present a novel approach to animate an entire tree from the motion of a few branches. Animating medium sized trees in wind presents an arduous situation because it is difficult to create the motion of flexible branches in turbulent wind flows. The animation of trees in wind is important in biology and ecology where understanding how trees move can inform us about the cultivation of trees and forests designed to withstand strong winds. Animation of trees in wind is also useful in the production of films and games. We use passive optical motion capture to record the motion of retro-reflective markers placed on some of the tree branch tips. These motion data are processed to remove noise, create motion traces, and add missing data. Given the processed motion data, we solve for wind velocity in a discrete formulation of aerodynamic drag and tree branch dynamics. This sparse collection of wind velocity samples is interpolated over the volume of the crown and enriched with a turbulence model to create a wind field that drives the animation of an entire tree crown. The wind field can also be used to drive the animation of other similar trees. In this process, 300 KB of data recorded from 10 s of tree motion can be processed and replayed in 16 s of computation time, not including rendering.