Elsevier

Computer-Aided Design

Volume 34, Issue 1, January 2002, Pages 19-26
Computer-Aided Design

Qualitative and quantitative comparisons of B-spline offset surface approximation methods

https://doi.org/10.1016/S0010-4485(00)00147-0Get rights and content

Abstract

Surface offset is one of the most useful operations in Computer Aided Geometric Design (CAGD). However, an implementation of this operator is not trivial primarily because the offset surface, in general, does not have the same representation as the original surface. Hence, it is difficult or impossible to represent an exact offset in a system with limited surface forms. For this reason, some CAGD surface offset operators produce results that are, at times, unsatisfactory. In this article, we discuss surface offset approximation methods in B-spline environment: both the original surfaces and their offsets are B-spline surfaces. This article summarizes research contained in the first named author's PhD thesis.

Section snippets

Surface offsetting

The first step in this research was to conduct a survey of known curve and surface offset methods.

A set of surface offset methods was compiled. (These methods will be discussed in detail in Section 4.) Some previously developed methods that were just for planar curves or just for cubic curves were extended to surfaces. Some of the studied methods preserve the original smoothness of the base surface, an important criterion in exterior surface modeling, and some do not. Below is a list of the

Surface definition

A tensor product B-spline surface has the form [1]:S(u,v)=(x(u,v),y(u,v),z(u,v))=i=0nj=0mfip(u)fjq(v)Pi,jwhere fip(u),fjq(v) are univariate B-spline basis functions of degrees p, q, respectively, and Pi,j are control points in R3. The (u, v) domain of this mapping is a rectangle in general.

Offset error analysis method

In the first phase of the research, the offset surface obtained from each offset method was compared to the actual offset computed on a subset of points from the original domain, using an analytical definition of the surface offset. While the above method provides an approximation of the error, it involves a large number of computations, and may miss some surface features.

In this research we were looking for an error approximation scheme that would apply to Bézier and to B-spline surfaces. We

Surface offset methods

Most of the tested surface offset algorithms require that the original surface be decomposed to a set of Bézier patches. Since in most cases the algorithms perform the offsetting on each patch, it is essential to know if the surface offset method preserves surface continuity between patches. Some of the tested methods do preserve up to C2 (curvature) continuity. Other methods might preserve only C1 (tangent) continuity, or only C0 (positional) continuity.

The following methods will be discussed

Qualitative comparison of surface offsetting methods

In a qualitative comparison of offset methods, the objective is to provide information on distinguishing attributes of these methods, their characteristics, or possession of some special qualities. The following list is a summary of such a comparison.

  • 1.

    Geometric methods usually underestimate the offset operator. During our research, we found the approximate offset obtained using geometric methods for a surface consisting of elliptic points will lie in a space between the original surface and the

Quantitative comparison of surface offsetting methods

In quantitative comparison, the major criterion is how efficiently each method approximates the offset surface given a prescribed tolerance. We compared the number of control points with respect to the accuracy of offset approximation, and the required offset distance. The following list is a summary of such a comparison.

  • 1.

    Overall, numeric methods (especially least squares methods) reached a preset error tolerance faster than geometric methods.

  • 2.

    Overall, the best numeric surface offset methods are

Comparison data

The summary of comparison charts [7] was organized in two groups: results for strict conditions (offset distance equal to 1 unit, and offset tolerance equal to 0.0001 units), and results for ‘loose’ conditions (offset distance equal to 2.5, and offset tolerance equal to 0.1).

The resulting charts illustrate the overall performance of all methods for a given surface and given conditions. Summary test results showed how many control points were required to offset a surface using all researched

Conclusion and recommendation

Throughout the conducted tests, we have observed the following consistent results:

  • 1.

    The Tiller–Hanson–Nachman method performs well in cases of surfaces with only elliptic points. This observation applies to most of the methods. Only approximation methods seem to be resistant to cases of rapidly changing curvature of the boundary curves.

  • 2.

    The biggest advantage of the geometric method is that it guarantees the same level of internal parametric continuity as the original surface. However, to maintain

References (9)

  • C. de Boor

    A practical guide to splines

    (1978)
  • S. Coquilart

    Computing offset of B-spline curves

    Computer-Aided Geometric Design

    (1987)
  • Elber G, Cohen E. Error bounded variable distance offset operator for free curves and surfaces. Salt Lake City, UT:...
  • Elber G, Lee I, Kim M. Qualitative and quantitative comparison of offset curve approximation methods. Department of...
There are more references available in the full text version of this article.

Cited by (18)

  • Volumes with piecewise quadratic medial surface transforms: Computation of boundaries and trimmed offsets

    2010, CAD Computer Aided Design
    Citation Excerpt :

    Hence, shape (curve, surface or volume) offsets and corresponding algorithms have been widely studied in computer-aided design (CAD) and manufacturing; see [1–8] and references therein for more details. However, free-form NURBS shapes do not possess rational offsets in general and thus suitable approximation techniques, often based on rational curves or surfaces of relatively low degree, are used; see, e.g., [9–14]. Nevertheless, these offset approximation techniques suffer from several shortcomings.

  • Computing exact rational offsets of quadratic triangular Bézier surface patches

    2008, CAD Computer Aided Design
    Citation Excerpt :

    More recently, offset computation of NURBS surfaces and of solids bounded by them has been studied in the paper series [14–16]. A qualitative and quantitative comparison of offset surface approximation techniques is given in [17]. Curve and surface modification in order to avoid local self-intersection is discussed in [18], and the detection and removal of self-intersections of offset curves and surfaces has been addressed in [19].

  • Triangular mesh offset for generalized cutter

    2005, CAD Computer Aided Design
  • Approximation of circular arcs and offset curves by Bézier curves of high degree

    2004, Journal of Computational and Applied Mathematics
View all citing articles on Scopus
View full text