This paper discusses how interval analysis can be used to solve a wide variety of problems in computer graphics. These problems include ray tracing, interference detection, polygonal decomposition of parametric surfaces, and CSG on solids bounded by parametric surfaces. Only two basic algorithms are required: SOLVE, which computes solutions to a system of constraints, and MINIMIZE, which computes the global minimum of a function, subject to a system of constraints. We present algorithms for SOLVE and MINIMIZE using interval analysis as the conceptual framework. Crucial to the technique is the creation of "inclusion functions" for each constraint and function to be minimized. Inclusion functions compute a bound on the range of a function, given a similar bound on its domain, allowing a branch and bound approach to constraint solution and constrained minimization. Inclusion functions also allow the MINIMIZE algorithm to compute global rather than local minima, unlike many other numerica...

Nonrobustness refers to qualitative or catastrophic failures in geometric algorithms arising from numerical errors. Section...

Many solid modeling construction techniques produce non-manifold r-sets (solids). With each non-manifold model N we can associate a family of manifold solid models that are infinitely close to N in the geometric sense. For polyhedral solids, each non-manifold edge of N with 2k incident faces will be replicated k times in any manifold model M of that family. Furthermore, some non-manifold vertices of N must also be replicated in M, possibly several times. M can be obtained by defining, in N, a single adjacent face TA(E,F) for each pair (E,F) that combines an edge E and an incident face F. The adjacency relation satisfies TA(E,TA(E,F))=F. The choice of the map A defines which vertices of N must be replicated in M and how many times. The resulting manifold representation of a non-manifold solid may be encoded using simpler and more compact data-structures, especially for triangulated model, and leads to simpler and more efficient algorithms, when it is used instead of a non-manifold repre...

Solid modelling and computational geometry are based on classical topology and geometry in which the basic predicates and operations, such as membership, subset inclusion, union and intersection, are not continuous and therefore not computable. But a sound computational framework for solids and geometry can only be built in a framework with computable predicates and operations. In practice, correctness of algorithms in computational geometry is usually proved using the unrealistic Real RAM machine model of computation, which allows comparison of real numbers, with the undesirable result that correct algorithms, when implemented, turn into unreliable programs. Here, we use a domaintheoretic approach to recursive analysis to develop the basis of an eective and realistic framework for solid modelling. This framework is equipped with a well-dened and realistic notion of computability which reects the observable properties of real solids. The basic predicates and operations o...

We present a system, ESOLID, that performs exact boundary evaluation of low-degree curved solids in reasonable amounts of time. ESOLID performs accurate Boolean operations using exact representations and exact computations throughout. The demands of exact computation require a different set of algorithms and efficiency improvements than those found in a traditional inexact floating point based modeler. We describe the system architecture, representations, and issues in implementing the algorithms. We also describe a number of techniques that increase the efficiency of the system based on lazy evaluation, use of floating point filters, arbitrary floating point arithmetic with error bounds, and lower dimensional formulation of subproblems. ESOLID has been used for boundary evaluation of many complex solids. These include both synthetic datasets and parts of a Bradley Fighting Vehicle designed using the BRL-CAD solid modeling system. It is shown that ESOLID can correctly evaluate the boundary of solids that are very hard to compute using a fixed-precision floating point modeler. In terms of performance, it is about an order of magnitude slower as compared to a floating point boundary evaluation system on most cases. 1

A new boundary evaluation method is presented. It is based on error-free Boolean operations on polyhedral solids. We describe, in detail, an intersection algorithm that handles, in a straightforward way, all the possible geometric cases. We also describe a general data structure that allows an unified storage of solid boundaries. The intersection algorithm always runs to completion, producing consistent solids from consistent operands. Numerical errors are handled at an algorithm independent level: an original exact arithmetic that performs only the necessary precise computations. Results from our implementation of this CSG solver are discussed.

We present efficient representations and algorithms for exact boundary computation on low degree sculptured CSG solids using exact arithmetic. Most of the previous work using exact arithmetic has been restricted to polyhedral models. In this paper, we generalize it to higher order objects, whose boundaries are composed of rational parametric surfaces. The use of exact arithmetic and representation guarantees that a geometric algorithm is numerically accurate and is likely to be required for perturbation techniques which handle degeneracies. We present efficient algorithms for computing the intersection curves of trimmed parametric surfaces, decomposing them into multiple components for e cient point location queries inside the trimmed regions, and computing the boundary of the resulting solid using topological information and component classification

An important part of solid modeling systems based on curved primitives is the representation and manipulation of algebraic plane curves with rational coefficients and points with algebraic coordinates. These objects are often approximated by floating-point numbers and spline curves, which are easy to manipulate, but are subject to accuracy and robustness problems. Exact computation can eliminate these numerical robustness problems, but efficient exact methods have not been available. Moreover, it is widely believe that exact arithmetic is impractical for manipulating curved primitives.

Geometric modelers are becoming faster and more powerful, but they still suffer from reliability problems because of floating point errors. Previous work in the field of robust geometric modeling tends to be problem specific and has proven hard to generalize. The approach described here is a general paradigm for handling the accuracy problem for a large set of geometric algorithms. This approach brings together ideas and techniques from interval arithmetic, constraint management, randomization, and algebraic geometry. It acknowledges that input values have tolerances, that objects within tolerance are equivalent, and that certain geometric singularities must be maintained because they reflect design intent or the laws of geometry. Our approach is systematic, and can be applied almost mechanically to the large domain of problems, that can be solved by algorithms using the operations +, -, * and /. The required theory and algorithms have been developed, and the viability of the concepts ...

We present a new model of geometric computation which supports the design of robust algorithms for exact real number input as well as for input with uncertainty, i.e. partial input. In this framework, we show that the convex hull of N computable real points in R^d is indeed computable. We provide a robust algorithm which, given any set of N partial inputs, i.e. N dyadic or rational rectangles, approximating these points, computes the partial convex hull in time O(N log N) in 2d and 3d. As the rectangles are refined to the N points, the sequence of partial convex hulls converges effectively both in the Hausdorff metric and the Lebesgue measure to the convex hull of the N points.