STATIC POLYHEDRON SIMPLIFICATION USING ERROR MEASUREMENTS

Polyhedral models of objects represent an efficient issue for reverse engineering applications and visualization purposes. Because digitizin g devices can produce large amounts of points, the polyhedral represen tation of objects may require a treatment of simplification to preserv e the efficiency of subsequent processes for machining or visualizatio n. Here, a new approach is introduced to reduce the number of nodes of a polyhedral model in accordance to an error criterion which can be e asily understood by the user. This criterion uses error zones assigned to each node of the initial polyhedron and guarantees that the simpli fied polyhedron intersects each of the error zones. This behaviour is equivalent to constraining the simplified polyhedron to be included in to the geometric envelope defined by all the error zones attached to t he initial polyhedron. The iterative treatments involved in the simpli fication process incorporate an inheritance mechanism to transfer the error zones from one iteration to the next one. Thus, an intuitive man ner to monitor the simplification process is set up. A node selection criterion based on curvature approximation has been included to sort t he nodes according to their probability of removal. A specific remeshi ng scheme has also been included to increase the robustness of the tre atment when face arrangements around a node generate distorted polygon shapes. Examples highlight the predictable behaviour of the algorithm . (C) 1997 Elsevier Science Ltd.