Generalized hidden surface removal

AbstractIn this paper we study the following generalization of the classical hidden surface removal problem: given a set S of objects, a view point and a point light source, compute which parts of the objects in S are visible, subdivided into parts that are lit and parts that are not lit.We prove tight bounds on the maximum combinatorial complexity of such views and give efficient output-sensitive algorithms to compute the views for three cases: (i) S consists of non-intersecting triangles, (ii) S consists of horizontal axis-parallel rectangles, (iii) S is the set of faces of a polyhedral terrain