Adaptive T-spline surface fitting to z-map models

Abstract

Surface fitting refers to the process of constructing a smooth representation for an object surface from a fairly large number of measured 3D data points. This paper presents an automatic algorithm to construct smooth parametric surfaces using T-splines from z-map data. The algorithm begins with a rough surface approximation and then progressively refines it in the regions where the approximation accuracy does not meet the requirement. The topology of the resulting T-spline surface is determined adaptively based on the local geometric character of the input data and the geometry of the control points is obtained by a least squares procedure. The advantage of the approach is that the resulting surface is C2 continuous and the refinement is essentially local, resulting in a small number of control points for the surface.