Page 18 - Geometric Modeling and Algebraic Geometry
P. 18

12     T. Dokken
                           q 2 (x, y, z)=0 of p 2 (s, t). Combining the parametric description of p 1 (u, v) and
                           the implicit representation q 2 (x, y, z)=0 of p 2 (s, t), we get q 2 (p 1 (u, v)) = 0. This
                           is a tensor product polynomial of bi-degree (54, 54). The intersection of two bicubic
                           patches is converted to finding the zero of

                                                           54 54

                                                                    i j
                                              q 2 (p 1 (u, v)) =  c i,j u v =0.
                                                           i=0j=0
                           This polynomial has 55 × 55 = 3025 monomials with corresponding coefficients,
                           describing an algebraic curve of total degree 108. This illustrates that the intersection
                           of seemingly simple surfaces can results in a very complex intersection topology. As
                           in the case of curves, the surfaces we consider in CAGD are bounded, and we are
                           interested in the solution only in a limited interval (u, v) ∈ [a, b] × [c, d].


                           1.5 Extend the use of algebraic geometry within CAD


                           The work within the GAIA project related to algebraic geometry and CAD has ad-
                           dressed three main topics:

                           •  Resultants are one of the traditional methods for exact implicitization of rational
                              parametric curves and surfaces. GAIA has produced some new results within this
                              classical research area.
                           •  Singularities in algebraic curves and surfaces are for understanding their geom-
                              etry and topology.
                           •  Classification is an old tradition in the field of Algebraic Geometry. It is a natural
                              starting point when trying to understand the geometry of algebraic objects.

                           Papers on CAGD and algebraic methods from the project are [8, 9, 32, 33, 34, 35,
                           41, 42, 44, 48, 49, 57].


                           1.5.1 Resultants
                           The objective has been to develop tools for constructing, manipulating and exploit-
                           ing implicit representations for parametric curves and surfaces based on resultant
                           computations. The work in GAIA has been divided into three parts:

                           •  A survey in four parts addressing:
                               1. A resultant approach to detecting intersecting curves in P .
                                                                               3
                               2. Implicitizing rational hypersurfaces using approximation complexes.
                               3. Using projection operators in Computer Aided Design.
                               4. The method of moving surfaces for the implicitization of rational parametric
                                 surface in P .
                                           3
                           •  A report addressing sparse/toric resultant, results when the number of monomials
                              is small compared to the number of possible monomials for polynomial of the
                              degree in question.
   13   14   15   16   17   18   19   20   21   22   23