I've written a piece of software that paints terrain with little "trees" as part of a research project here at Oregon State University called GNNViz. The trees density is determined by real tree density data (represented by grayscale tree density masks.
I was thinking it would be interesting to have it work in L3DT as a plugin, either using the attributes map to determine where trees go, or an external tree density mask (which is what I have). The current program is written in C using the "GD" graphics library, primarily to just do image merges of little tree pictures on top of a base image.
I see that the API has pixel-level get/put functions. Are they an efficient way to copy/merge sub-images into a larger terrain texture image when there are literally hundreds of thousands of little tree images to be placed?
Here's a small (512x512) version of what kind of image I'm talking about....
Here is a sub-sample of the full scale image...
You can view the much larger and much finer detail version at http://www.fsl.orst.edu/lemma/gnnviz/im ... posite.jpg