This paper presents a new method for voxelization of solid objects containing sharp details. Voxelization is a sampling process that transforms a continuously defined object into a discrete one represented as a voxel field. The voxel field can be used for rendering or other purposes, which often involve a reconstruction of a continuous approximation of the original object. Objects to be voxelized need to fulfill certain representability conditions; otherwise, disturbing artifacts appear during reconstruction. The method proposed here extends the traditional distance-based voxelization by an a-priori detection of sharp object details and their subsequent modification in such a way that the resulting object to be voxelized fulfills the representability conditions. The resulting discrete objects are represented by means of truncated (i.e., narrow-band) distance fields, which provide reduction of memory requirements and further processing by level set techniques. This approach is exemplified by two classes of solid objects that normally contain such sharp details: implicit solids and solids resulting from CSG operations. In both cases, the sharp details are rounded to a specific curvature dictated by the sampling distance.

We present a new technique for CSG operations with voxelized geometric objects, which are represented by truncated discrete distance fields supplemented by additional information about the surface normal. The technique removes artifacts of straightforward volumetric CSG operations by taking into account conditions for object representability, according to which sharp details are not correctly representable in discrete distance fields. The proposed technique solves this shortcoming by rounding edges and other sharp details. It works at the voxel level without the necessity for reconstruction of continuous object models.