Approximate Implicitization of Planar Parametric Curves using Quadratic Splines1)
Xiao-Shan Gao,Ming Li
2003-01-01
Abstract:A quadratic Bezier spline with G1-continuity is given to approximate a planar parametric curve. In its construction, the parametric curve is first divided into several segments. A rational quadratic Bezier curve is then applied to approximate each segment and recursive subdivision for the segment may be used if the approximation error is out of the given range. With this method, we not only keep the shape but also the direction of the parametric curve with simple computation using quadratic curves. High accuracy of the approximation is also achieved with relatively small number of conics. Parametric and implicit representations are two main forms for planar curves in CAGD. Each of the two representations has its own advantages. With the parametric representation, x = x(t)=d(t);y = y(t)=d(t), where x(t);y(t);d(t) are often polynomials (i.e. the rational form), it is easy to get every point in the curve and plot it. While for the implicit form f(x;y) = 0, it's easy to determine whether a point is on the curve. The availability of both often results in simpler computation. For example, with the two forms available, we can get the intersection of two planar curves by solving a one-variable equation. For any rational parametric curves, we can always convert it into implicit form. The process is called implicitization. While the converse, called parametrization, is not always true. The implicitization techniques include some classic algebraic geometry theories, such as the resultant based method, the Groebner basis method, the characteristic set method, and others such as moving lines and surfaces. For a survey, please consult (13). However, it is not necessary to give an exact implicit form for a parametric curve. This is in part due to the fact that the implicitization always involves relatively complicated process and the result of the implicit form might have large number of coecients. Another