This thesis addresses the thinning of a point cloud for the purpose of numerical simulations. From input CSV data, obtained using the DIC method, a mesh is generated over the reduced set of points using three different strategies. The aim is to reduce the number of points while preserving the geometric distribution of the original cloud and decreasing numerical complexity. The results show that uniform quadrilateral gridding achieves the most efficient thinning and meshing, while triangle-to-quadrilateral merging, with an enhanced algorithm, also showed potential.
|