Numerical Proper Reparametrization of Space Curves and Surfaces. (November 2019)

• Record Type: Journal Article
• Authors: Shen, Li-Yong; Pérez-Díaz, Sonia; Yang, Zhengfeng
• Is Part Of: Computer aided design. Volume 116(2019)
• Publication Date: 2019-11
• Subjects: Numerical/symbolic reparametrization -- Space curve -- Rational surface -- Approximately improper/proper; Computer-aided design -- Periodicals; Engineering design -- Data processing -- Periodicals; Computer graphics -- Periodicals; Conception technique -- Informatique -- Périodiques; Infographie -- Périodiques; Computer graphics; Engineering design -- Data processing; Periodicals; Electronic journals; 620.00420285
• DOI: 10.1016/j.cad.2019.07.001
• Languages: English
• ISSNs: 0010-4485

Abstract

Abstract: Simplifying rational parametrizations of surfaces is a basic problem in CAD (computer-aided design). Reducing their tracing index, called proper reparametrization, is an important simplification way. Most existing proper reparametrization work is symbolic. Yet in practical environments the surfaces are usually given with perturbed coefficients hence need a numerical technique of reparametrization considering the intrinsic properness of the perturbed surfaces. We present algorithms for reparametrizing a numerically rational space curve or surface. First, we provide an efficient way to find a parametric support transformation and compute a reparametrization with proper parametric support. Second, we develop a numerical algorithm to further reduce the tracing index, where numerical techniques such as sparse interpolation and approximated GCD computations are involved. We finally provide the error bound between the given rational curve/surface and our reparametrization result. Graphical abstract: Highlights: We introduce the numerical proper reparametrization of space curves and surfaces. Give a new algorithm with linear complexity to find the transformations for improper supports. The numerical algorithms are designed for approximately proper reparametrizations. The error between the numerical curve/surface and their reparametrizations are studied.