Ricostruzione adattativa di dati sparsi mediante funzioni a simmetria radiale
Tuesday 20th December 2011
This work deals with the reconstruction of implicit surfaces from 3D scattered data. Our method is based on the radial basis function (RBF) approach for the multivariate interpolation. Starting from an existing algorithm that establishes a multi-level reconstruction by using compactly supported radial basis functions, we developed a new procedure in order to get an adaptative reconstruction at each scale. The numerical validation shows the new procedure to be attractive for its capability of providing more accurate reconstructions with less degrees of freedom compared with the original algorithm, the computational time being almost the same. Some limitations due to the RBF interpolation itself are still present, in particular in the reconstruction of noisy clouds of points as well as processing sharp featured clouds. Enhancing performances in such cases will be the object of the future development of this research.