This is not an automatic system, but it allows control the number of nodes for mesh imaging efficiency. This method may be controversial because it ignores the abstraction of contours on which so many users rely. Replicating contours produces too many meaningless nodes.
I suggest this:
Place the JPEG image in an Archicad plan view at the correct size. (I use this method on vector drawings, too)
Draw a suitable grid over the image. The grid interval depends on the finesse you need and the steepness of the site.
Build the mesh by placing mesh nodes corresponding in height to the nearest contour line at each grid intersection. Always use the lower contour.
Unimportant areas can have fewer nodes: every other grid intersection.
Important areas like rock outcroppings should have many more nodes to allow for sharp definition. Here you would closely follow a contour. Road edges, too, are important. You might link nodes across and along road surfaces to control road slope appearance.
This method provides a quick and polygon-efficient mesh and takes only a few minutes to make. A little work invested at the start speeds up rendering and subsequent 3D views.
Attached is an example of a quick rendering using this method for mesh AND also using a jpeg image with roadways stretched over as a material.
Dwight Atkinson
