Maps from Tracks?
I'm happy to have found this community after spending too much time lurking about traditional GIS folk while learning what I could for a cartography-based art/design project.
The project is to create maps that are constructed essentially from real world analytics derived from GPS tracks. These tracks are meant to be used to construct "fantasy" terrains that correlate to where I've been and how often I've been there.
I've tried a couple of approaches, one of which results in these kinds of maps generated from GPS tracks:
So, the idea is that I only see where I have been in the maps, and larger islands/continents for places that I have been more often (home/studio). These maps above basically show a couple of months of spending time around home and thereabouts, and driving up to the studio where I work.
This is satisfactory, but I'd like to go a step further and take advantage of some other tools to enhance the maps. For instance, I'd like to generate fractal terrains. I've tried creating height maps for Fractal Terrain, but that hasn't gone particularly well, I think because FT expects much more detail than I am producing, and probably some other reasons that I haven't figured out yet.
What I would expect to do in an ideal world is to run my GPS tracks (GPX data) through a pre-processor that I'd probably construct and use the output as source for some tool that would build-up the baseline fractal terrain that I could then muck with to bring it to life. But, I'd like the baseline terrain to be generative, based on the GPX data.
I've looked at FT and now I'm eyeballing World Machine.
Anyone have suggestions of other tools or approaches?
Presumably you have a lot of tracks that are in this vector format and that these tracks basically dont backtrack on themselves on a per track basis but do a lot of overlap between tracks. So you could draw these tracks with a thick but blurry line (gaussian profile) and then add them all up and normalize the result to the input for your terrain app. I suspect that your tracks are made up of GPS samples so putting splats down and accumulating them might be easier than using lines between the samples. Waldronate said that FT can export raw data so I presume it can import it. So if you have a rect array of floats then you should be able to import that and process it more from there - or use Wilbur and then transfer to FT. If your programming this anyway then maybe get a library for height maps - I think the GDAL is an open source thing which can do all sorts of terrain format stuff like that.
Yep, tracks, which are basically collections of latitude/longitudes/altitudes and some other meta data. I've got a pre-processor that'll take this data, correlate it with some other stuff and create what are roughly heat maps. I can assign color gradient files, so I could create them as greyscale image files. If tracks run over each other over time, which they do, they "heat up", and heat up surrounding locations as well.
So, that's the start.
Over the weekend, I wrote an exporter that generates DEM files that FT can import. (If there are other binary formats FT can read, I'd be eager to hear more about it! I'm a newbie to FT.) But, sadly — I don't think there's enough detail there as it is. FT runs across an entire world (-180 — 180 / -90 — 90) and my footprint on the world is limited to a small swath of Southern California, with jaunts to New York City and a few cities in Europe. (It's humbling to realize how much of the world I'll never see, which is one of the motivations for creating these maps.)
I think I need to find a way to get pre-process my data differently, to add more detail, or coverage to my maps.
Or, perhaps there's a different way to have FT embrace my data?
Tags for this Thread