"Implementação numérica do método Level Set para propagação de curvas e superfícies"

Abstract

In this dissertation, we present a powerful numerical technique known as Level Set Method for computing and analyzing moving fronts in different physical settings. The method formulated by Osher and Sethian [1] is based on the following idea: a curve (or surface) is embedded as the zero level set of a higher-dimensional function Φ (called level set function). Then, we can link the evolution of this function Φ to the propagation of the curve itself through a time-dependent initial value problem. At any time, the curve is given by the zero level set of the time-dependent level set function Φ. The evolution of the level set function Φ is described by a Hamilton-Jacobi type partial differential equation, which can be discretised by the use of accurate methods for hyperbolic equations. As a result, the Level Set Method is able to track complex curves that can develop large spikes, sharp corners or change its topology as they evolve. Because of its versatility and efficacy, this numerical technique has found applications in a large number of areas, including fluid mechanics, image processing and computer vision, crystal growth, computational geometry and materials science. Particularly, the aim of this dissertation has been to understand the fundamentals of Level Set Method and its final goal is compute the motion of bondaries in crystal growth using this numerical model.