- Joined
- Jun 22, 2008
- Messages
- 6,368
- Reaction score
- 0
- Points
- 0
For the last few weeks, I've been trying to create speculative planetary bodies. I know that sounds like a tall order; that's because it is.
Mainly I've been focused on a world with oceans, continents and tectonic activity, similar to Earth. It is quite a complex system to understand; plate boundaries transform over time, plates are constantly rifting, subducting, etc. Everything needs a consistent history.
Eventually, I felt like subducting my head through my desk...
So I decided to drop everything that I was doing (an example of why I'm awesome at setting a goal and completing it) and decided to make a Moon or Mercury-like world instead.
It sounded so simple at the time. For good reason; there's no tectonic activity on the Moon, and volcanic activity seems fairly limited (at least in comparison to Mars). The major features on the surface are formed by impact events- simple excavations of varying size, randomly distributed across the surface, layered over eachother in billions of years of bombardment.
So the method of creating such a world dawned on me; first create a greyscale heightmap of the object in question, by layering 'craters' (variations in shade, representing variations in altitude) of varying size over the image, guided by random points (to be generated by a random noise function or similar). After a specific set of iterations (representing the early history of the planet, when impacts would have been common), the average size of the craters would be reduced. In addition, throughout the process alterations would be made to account for phenomena such as volcanic eruptions and general wearing-down of features due to general [ame="http://en.wikipedia.org/wiki/Impact_gardening"]impact gardening[/ame].
Finally, a visual map of the object would be completed- albedo and colour features- based on factors such as craters (which turn up fresh regolith and create visually impressive ray systems) and old lava fields (like lunar maria).
I did not pay much concern to the issue of polar distortion due to the map projection. There are however methods to rectify this issue.
The problem I encountered, of course, is that image editors do not quite work in the same way as planets. Problem one is that using a specific image editing tool (primarily the dodge/burn tool, but also the normal 'paintbrush' tool at non-total opacity) when creating overlapping craters.
A common phenomenon is a younger crater superpositioned over an older one. The depths of each crater can be roughly similar, and the rim of the younger crater is above the floor of the older one, disrupting the outline of the older crater.
In my case, however, overlapping two 'craters' creates a lens-shaped darker area (which would represent a 'depression') between the two- this is a feature completely antithetical to reality.
Another paintbrush-tool-method (forgive my terminology; the proper terms confuse me), which (as far as I can understand) 'pastes' an image onto the picture, creates a crater-crater relationship quite similar to that in reality, but produces a new fatal flaw; since the same image is added with every brush 'stroke', every crater is the same depth, regardless of the altitude of the area it impacted, or the size of the crater in question! This too, of course, is antithetical to reality.
Here is a WIP heightmap in all of its unglory;
I'm not sure how to progress now. Undoubtedly my method is simplistic, and there are clearly certain aspects of lunar or mercurian geology that I'm missing- like rille, scarps, volcanic domes, etc, or the effects of distributions of elements and minerals on colouration and albedo. But I can't move on to focus on those things if I can't get the basic features of lunar-like geology right- and those features are impact craters and basins.
Mainly I've been focused on a world with oceans, continents and tectonic activity, similar to Earth. It is quite a complex system to understand; plate boundaries transform over time, plates are constantly rifting, subducting, etc. Everything needs a consistent history.
Eventually, I felt like subducting my head through my desk...
So I decided to drop everything that I was doing (an example of why I'm awesome at setting a goal and completing it) and decided to make a Moon or Mercury-like world instead.
It sounded so simple at the time. For good reason; there's no tectonic activity on the Moon, and volcanic activity seems fairly limited (at least in comparison to Mars). The major features on the surface are formed by impact events- simple excavations of varying size, randomly distributed across the surface, layered over eachother in billions of years of bombardment.
So the method of creating such a world dawned on me; first create a greyscale heightmap of the object in question, by layering 'craters' (variations in shade, representing variations in altitude) of varying size over the image, guided by random points (to be generated by a random noise function or similar). After a specific set of iterations (representing the early history of the planet, when impacts would have been common), the average size of the craters would be reduced. In addition, throughout the process alterations would be made to account for phenomena such as volcanic eruptions and general wearing-down of features due to general [ame="http://en.wikipedia.org/wiki/Impact_gardening"]impact gardening[/ame].
Finally, a visual map of the object would be completed- albedo and colour features- based on factors such as craters (which turn up fresh regolith and create visually impressive ray systems) and old lava fields (like lunar maria).
I did not pay much concern to the issue of polar distortion due to the map projection. There are however methods to rectify this issue.
The problem I encountered, of course, is that image editors do not quite work in the same way as planets. Problem one is that using a specific image editing tool (primarily the dodge/burn tool, but also the normal 'paintbrush' tool at non-total opacity) when creating overlapping craters.
A common phenomenon is a younger crater superpositioned over an older one. The depths of each crater can be roughly similar, and the rim of the younger crater is above the floor of the older one, disrupting the outline of the older crater.
In my case, however, overlapping two 'craters' creates a lens-shaped darker area (which would represent a 'depression') between the two- this is a feature completely antithetical to reality.
Another paintbrush-tool-method (forgive my terminology; the proper terms confuse me), which (as far as I can understand) 'pastes' an image onto the picture, creates a crater-crater relationship quite similar to that in reality, but produces a new fatal flaw; since the same image is added with every brush 'stroke', every crater is the same depth, regardless of the altitude of the area it impacted, or the size of the crater in question! This too, of course, is antithetical to reality.
Here is a WIP heightmap in all of its unglory;
I'm not sure how to progress now. Undoubtedly my method is simplistic, and there are clearly certain aspects of lunar or mercurian geology that I'm missing- like rille, scarps, volcanic domes, etc, or the effects of distributions of elements and minerals on colouration and albedo. But I can't move on to focus on those things if I can't get the basic features of lunar-like geology right- and those features are impact craters and basins.
Last edited: