3D terrain - proof of concept

I had a stab at this when I first started dabbling with Processing and didn't get very far...  I've obviously improved substantially since then as it only took me an afternoon to get my original code working and to then get it to read height and colour data from images .  (IIRC it took me somewhat longer to produce the original code that didn't even work!)

I've no idea if this is the best approach to this kind of project - I'd initially considered using arrays to store height and colour data but I'm assuming that using images in this way is at least no less efficient...  If nothing else it's a heck of a lot easier to generate the data (I tested the array version by randomly populating the arrays!)

Next I'd quite like to add objects to the terrain - I think I'll start with some trees - and I suspect it's at this point that I'll discover whether this approach is actually practical...

Any feedback gratefully received!

Thanks

I guess I didn't phrase my comments too well: I did know that greyscale maps were often used for this purpose, but what I wasn't sure of was how they were converted to 3D.  I did wonder whether they'd be used to create a complete 3D mesh and then some other method was used to mask the sections of the mesh not on screen, but that didn't seem to make a lot of sense.

I have discovered and now fixed one slight issue (though not yet uploaded it): namely that the number of 'tiles' is actually less than the number of points, so the index reference for the colour map is slightly off.  You don't see this in the demo I've posted, but I eventually want to get it to wrap around at which point there would be obvious seams...  I'm still trying to figure out how to avoid outOfBounds errors on wrapping at the moment though; but need a break first!

I take your point on using proper image textures, but I rather like the retro look - it takes me back to my younger days.  Actually my inspiration for this was Zarch later released as Virus.  Back in the day people weren't scared to release games with fiendishly difficult control-schemes

Out of curiosity I'd be interested in knowing the relative benefits/costs (i.e. regarding memory and processor load) of using an image to store and process array data, rather than using an actual array.  It also strikes me that I could potentially use one channel of the terrain map to store height information and use the other two channels for other information - or is that asking for trouble

Hi,

these kinds of questions are a common part of raycasting actually.  i would suggest checking out

http://www.flipcode.com/archives/Realtime_Voxel_Landscape_Engines-Part_1_Introduction.shtml

and

http://www.advsys.net/ken/voxlap/voxlap03.htm