|
(0021394)
|
|
Ken Moreland
|
|
2010-07-14 17:52
|
|
If you decide to implement contours on cell data, what algorithm do you plan to use do it? The most common approach is to run cell to points and then contour. This is a quick and dirty approach but technically wrong. One of the biggest problems is that it does not preserve maxima and minima.
I advocate using a simple algorithm that extracts all faces that are the shared interface between two cells that have field values on opposite sides of an isovalue. Yes, the resulting surface will be boxy, but that is not necessarily bad. After all, the pseudocoloring for cell data is boxy, so why shouldn't contours be? |
|