Extracting geometric features from a virtual environment

In the fields of computer aided design (CAD) and virtual reality (VR), there have been various approaches in converting geometric models from a CAD environment to a VR environment for virtual prototyping purposes. Comparatively, little work has been published on the inverse modelling issues from VR to CAD. Robust algorithms on inverse modelling will facilitate model updating if an object is directly modified in a virtual environment (VE). This paper presents algorithms for constructing a high-level CAD model from a polygonal model found in a virtual environment. The faceted model is first uniformly sliced in three-orthogonal directions and a contour-based surface model is created. Boundary points are then identified from the contour model. A network of boundary curves is further constructed with the vertices of the neighboring polygons of the original VR model derived from the previously identified boundary points. The boundary curves partition the entire surface of the object into individual surface areas and form the topology of the object. The type of an individual surface is further identified. A concise surface description is then reconstructed for each of the surface areas from the corresponding vertices falling inside the topological domain. Some examples are presented to illustrate the developed algorithms.