6.2.2 The Delaunay Triangulation



next up previous contents index
Next: 6.2.3 Problem Decomposition and Up: 6.2 Architectural Design Previous: 6.2.1 Related Problems and

6.2.2 The Delaunay Triangulation

   

Delaunay triangulation, Voronoi tessellation, Lawson's triangulation, the construction of Thiessen or Dirichlet regions all denote the same problem. This problem is so elementary that it can be found in virtually every scientific and engineering discipline that uses a conception of space. Simply statedgif, the problem is as follows (note that the formulation is independent of the spatial dimensionality).

   

 

  The Delaunay graph (the Delaunay triangulation in the two-dimensional case) is easily obtained and closely related to the definition above.

  

From this definition one easily concludes that a Delaunay triangle is formed between the three points , , and if and only if the three Voronoi edges (which are the perpendicular bisectors of the edges of the Delaunay triangle) exist and hence we obtain the following criterion:

and this is the circumcenter of the triangle and is the radius of the circumcircle. This means that the triangle is a Delaunay triangle if and only if there is no other point of inside its circumcircle.

Speaking more intuitively, in the two-dimensional case the Voronoi regions are those areas where every point is closer to one of the grid points than to any other grid point. The (straight line) borders of the Voronoi regions are called Voronoi edges and are the set of points of equal distance to two nearest grid points. Every grid point lies within exactly one associated Voronoi region. Two grid points are next neighbors and are connected by a Delaunay edge when their Voronoi polygons touch each other, when they share a common Voronoi edge. The Voronoi edges are symmetry axes for next-neighbor grid point pairs.

Many different direct and indirect construction techniques for the Delaunay triangulation are available and it is beyond the scope of this section to give an extensive overview. A good introduction and explanation of the basic properties is given in [95].          

The following nomenclature is adopted for describing the algorithms and data structures in the subsequent sections. A (directedgif) graph consists of a set of nodes and a set of edges , , where an edge is an ordered pair of nodes. The edge is said to be incident to and the node is incident to . The nodes and are then adjacent (via the edge ). The edge is adjacent to , as and share the common node .

The given points (the in Definition 6.3) of the point cloud are called Delaunay nodes, as they are the nodes of the Delaunay graph. The Delaunay edges are the edges of the Delaunay graph and they connect all Delaunay nodes.

  
Figure 6.1: Grid point cloud

Figure 6.1 shows a set of Delaunay nodes that are triangulated to form the Delaunay graph shown in Figure 6.2.

  
Figure 6.2: Delaunay graph for the grid point cloud

The Voronoi graph consists of Voronoi nodes (these are the circumcenters of all Delaunay triangles) and Voronoi edges (these are the sets of all that adhere to Equation 6.1). The Voronoi graph of the point cloud in Figure 6.1 is shown in Figure 6.3.

  
Figure 6.3: Voronoi graph for the grid point cloud

The following important dualities between the Voronoi graph and the Delaunay graph can be observed.

 
Table 6.1: Duality relations between Voronoi and Delaunay Graph

Note that in an unconstrained Voronoi tessellation, the boundary of the triangulated domain is always a convex polygon and the Delaunay nodes on the boundary correspond to open Voronoi polygons. Open Voronoi polygons contain two infinite Voronoi edges, which are dual to Delaunay boundary edges.



next up previous contents index
Next: 6.2.3 Problem Decomposition and Up: 6.2 Architectural Design Previous: 6.2.1 Related Problems and



Martin Stiftinger
Thu Oct 13 13:51:43 MET 1994