Technically, the creation of these images is not very complex. The first tool are my eyes, to look at the real world. Then I use a text editor (currently SciTE) to type the description of the objects using the POV-Ray language. This text file is used as input for the POV-Ray raytracer, to render the scene description into a graphic file. In the making process, I also use The Gimp to create or retouch height and material maps used into the scene. All that is done curently with an AMD 64 X2 5600+, running Linux Fedora 9, and using the latest versions of POV-Ray and MegaPOV, compiled from sources for this platform.
At first, it can look very difficult to type directly in a text file the 3D description of the scene to be rendered. But when learned, it is the only "true 3d" interface (you enter data directly in 3D form), and also the only available to people with low hand-eye coordination (as me). It has many limitations, but also some advantages, when compared with the typical CAD/modelling package. Anyhow, I just started to use also a modeller to create some objects. I never got used to any other modeller, but Wings3D is really different, very intuitive and simple to use. Surely my scenes will benefit greatly, at least in respect to the shapes.
For my work methods... better don't ask. I'm not very conscious of what I do when I do it. Latter, I understand it and try to clean a bit the code. So, the best I can do is to share it, rather than explaining techniques that consist, mostly, on "trial and error" and "muddling along".
So, at the image galleries you can find the links to the source code for the images, if it is available. If there is not the code you want, let me know and I will put a link soon. Also, in this section, you can find some tips & tricks, the results of my experiments, and also a section dedicated to the LightSys macros.