Next: Homework: Due Week 3
Up: Week 2: Mathematics and
Previous: Theory
Subsections
Practice: Mesh and Display Issues
Read CW [10] Chapter 8 and pages 136-152 in Chapter 6.
-
- Order and Continuity
-
- Problems and benefits with sampling over patches.
-
- Problems and benefits with sampling at vertices.
-
- Common mesh artifacts.
-
- Density
-
- Shape
-
- Aspect ratio
-
- Grading (Mesh size varies smoothly)
-
- Conformance (Vertices line up across boundaries)
We are discussing a priori meshing (meshing that happens before
the solution takes place) but all of these points apply to adaptive
meshing as well.
In real radiosity systems, the input data comes from a variety of
sources. This data is usually not ideal, and may be far from it.
The input geometry should be pre-processed in order to get it ready for
meshing. This involves several steps:
-
- Merge coplanar, adjacent surfaces.
-
- Remove unnecessary vertices
-
- Worry about degenerate geometry.
-
- Worry about numeric problems.
After preprocessing, the initial mesh can be created. For now, the
initial mesh is also the final mesh, so it must be created fine enough
to capture all interesting detail. Eventually, the initial mesh will
be subdivided by the radiosity system. Depending upon how this is
done, the initial mesh should be created as large as possible, so that
the number of mesh elements is as small as possible.
-
Grid Superposition
- Superimposes a grid over the surface. Has
problems at edges (Used by Baum etal. [4]).
-
Template Mapping
- Mesh in parameter space.
-
Multiblocking
- Chop geometry into simpler pieces and create
the mesh on those. This includes D0 discontinuity mesh and
Campbell's BSP tree meshing.
Ronchamp: A Case Study [18]
Making Radiosity Usable: Automatic Preprocessing and Meshing
Techniques for the Generation of Accurate Radiosity Solutions [4]
Discontinuity Meshing [23]
BSP trees for discontinuities [6]
Next: Homework: Due Week 3
Up: Week 2: Mathematics and
Previous: Theory
Comments: Brian Edward Smits
1998-06-08