An idea off the top of my head, if you rendered your scene with a 
black/white gradient texture, to create a depth map, the depth map could 
easily be used to locate edges with a post process. Those edges could then 
either be directly applied to the scene, rendered with high ambient 
textures, or converted to cylinder/sphere line structures and placed in 
front of the camera to be rendered along with a high ambient scene.

Perhaps this process could be duplicated entirely in pov code using some of 
those crazy trace features. A high pass filter could then be programmed 
into the depth values array, and then threshold applied, vectors sussed 
out, and then, as parsing completes, cylinder structures placed in the 
proper place.

A bit of a complicated solution, but it could work.

Post a reply to this message