I Can See Your House From Here
April 21, 2010 10:22 AM   Subscribe

grassrootsmapping.org also showed up in this AskMe a couple of days ago.
posted by le morte de bea arthur at 10:32 AM on April 21, 2010

These aren't actually maps, though. These are just aerial photographs (although cool ones). A map is supposed to give you more information about a thing or a place by abstracting away unnecessary stuff and thus making other things easier to find and easier to see (by clearing away the things that are covering them). These pictures just show you how a place looks like from a balloon; it's quite difficult to find any useful information there.

Grassroots mapping is a great idea, but these photos are just the beginning, the raw input; there's still some work to do before they can be called maps.
posted by daniel_charms at 10:39 AM on April 21, 2010 [2 favorites]

Part of me that hoped that grassroots maps were cross section diagrams of the root structures of various plants....
posted by Hoenikker at 10:52 AM on April 21, 2010

Oh, and the kids these days, they have it easy. When I was a kid, we had to make a map in the geography class and we had to make do with our legs and a compass (and pencil + plotting paper). We were lucky if we got to use a measuring tape to check our measurements!
posted by daniel_charms at 10:57 AM on April 21, 2010

I agree with daniel_charms, I was all "where them maps at?" This is remote sensing. Wait til they get all IDRISI on that shit, then you'll have some maps. But sure, still cool.
posted by desjardins at 11:13 AM on April 21, 2010 [1 favorite]

I do wish they would go the next step and annotate the photos. To me, that could me more of a map. I was disappointed that the Flickr images contained no notes. Perhaps that's an envisioned next step. Still, I like the idea and the collaborative nature of the project. I'm sure it will evolve.
posted by Miko at 11:22 AM on April 21, 2010

Grassroots mapping is a great idea...

posted by DU at 11:54 AM on April 21, 2010 [1 favorite]

this has a lot of potential for classrooms, especially for disadvantaged students (if you can afford the field trip, that is). didn't we have an awesome thread here in the last several months on using GIS, utility databases & property appraiser dbs to show social injustice at the local government level?
posted by toodleydoodley at 1:09 PM on April 21, 2010

These aren't actually maps, though....

They can become maps though. Warren has developed Cartagen which you can see combined with the aerials, if you scroll down here. It's a way of generating vector maps using HTML, which is pretty cool
posted by a womble is an active kind of sloth at 1:24 PM on April 21, 2010

womble: Maybe I'm not looking right, but the only thing I see that Cartagen does is warp and stitch photos together using a pre-existing vector map as a basis. Which is pretty cool in itself, but not what you seem to think it is.
posted by daniel_charms at 1:57 PM on April 21, 2010

Just to clarify, I realize that the 'human' is doing the image processing, and identifying the features rather than a computer. I would consider that to be a map, though.
posted by a womble is an active kind of sloth at 1:59 PM on April 21, 2010

Awesome to see that this is getting some attention. I've hung out with Jeff a number of times and the level of theory and implementation here is amazing.

Given the discussion afterward, it might be useful to clarify a few things:

These are definitely maps. Some of the blog posts have a photo that looks like a photo, but the end goal is to georectify these images and have them be super-high-res kite maps of small areas. There is no definition at all of maps that requires them to simplify information, and satellite maps are commonly referred to as maps. Referring to it as remote sensing is like referring to the tools as GIS instead of software: there's a greater loss in popular understanding than there is a gain in specificity.

To point out OpenStreetMap as prior art is kind of obvious and misses a huge point here: OpenStreetMap is seeded by Yahoo's aerial imagery, that it very kindly provides for their tracing efforts. That's great, but there's obviously an impurity in this plan; the open source process begins with a closed source that's liberally licensed. Kite mapping and other efforts give the opportunity for a full-stack open mapping effort, and for a huge new level in the time it takes to map an area and the resolution to which it can be mapped. And the tools, like GDAL, ossim, etc., which are developed for small swaths, and the web interfaces that are being worked out here, can scale to other scenarios (like raw satellite data)

And cartagen is a very interesting development as well: it renders maps from OpenStreetMap-formatted source data in the browser, which means that there's again a higher level of flexibility here than tile-rendered maps and a more pure process. There's also Cartagen Knitter, which is a very cool attempt at mainstreaming the idea of rubbersheeting to make map rectification something that the bored masses can do.
posted by tmcw at 5:04 PM on April 21, 2010

There is no definition at all of maps that requires them to simplify information...

No, that is exactly the definition of a map. Borges has a story about a "map" that is exactly the size of the country it portrays and it's ridiculous. If it isn't simplifying somehow it's just reality, not a map.

OpenStreetMap is seeded by Yahoo's aerial imagery, that it very kindly provides for their tracing efforts.

This is one of many sources for OSM's maps. Most of their data is from public sources, including people walking around with GPS units and/or hand editing maps of areas they personally know.
posted by DU at 5:44 PM on April 21, 2010

The map is not aerial photos of the territory.
posted by mendel at 7:18 PM on April 21, 2010

tmcw: “OpenStreetMap is seeded by Yahoo's aerial imagery, that it very kindly provides for their tracing efforts. That's great, but there's obviously an impurity in this plan; the open source process begins with a closed source that's liberally licensed.

I think that's being pretty dismissive of OSM. My understanding is that when OSM was getting started, there was no assumption that they'd be able to use Yahoo's high-res imagery. The idea was to use publicly-available satellite imagery (NASA stuff) where possible, and then build up the street-level data from GPS tracks.

OSM's software reflects this; you can drive (or walk, or bike, or whatever) with a GPS unit running, and then upload the tracks into OSM. Even if you don't do anything else, just doing that helps the effort in a way, because other people can view your tracks and use them to create the actual vector maps.

At some point OSM got access to all of Yahoo's satellite data via their API, which has been a boon to the project; you can easily fire up their tool and sanity-check the vector map against the real world, fix errors, etc., without any GPS tracks. It makes the process much more approachable. But in the absence of it — or in areas where the Yahoo imagery doesn't exist or isn't very good — you can still work from GPS tracks. I've done this for things that just aren't visible on the sat photos: roads with lots of treecover, bike trails, walking paths, etc.

All that said, I do agree with your overall point that kite imagery could do a lot for bottom-up mapping projects, and it would fit pretty well into OSM, if it's being done in areas that aren't already well-covered by high-resolution aerial photos from Yahoo. If there's already sub-6" resolution imagery via Yahoo's API, liberally licensed, then it seems a bit like wheel reinvention. But if the only liberally licensed imagery of an area is decades-old Landsat crap, then it's pretty awesome and I can imagine that a lot of people — including OSM — would be really eager to get their hands on it.

I think the whole grassroots mapping thing is pretty neat, and I'm a map junkie so as far as I'm concerned, more and better maps are always a Good Thing, but I also think it's important for people to try and coordinate their efforts and work together. It's pretty silly to spend hours tracing vectors on a balloon image if the area in the image has already been mapped in OSM, and you could instead update the OSM data, or at least use it as a starting point. There are a lot of applications, most of which we can't really even imagine now, that become possible when you have very high-quality, freely-available map data. The more people we can get, and the more approaches we can use to gather that data, the better — but it would be disappointing if the overall possibilities were diminished because people developed a case of NIH syndrome.

I'm not saying that's what's happening, but just that it's something that should be actively avoided as more and more people get into cartography. We can do a lot more together than if we're all working separately.
posted by Kadin2048 at 7:23 AM on April 22, 2010 [2 favorites]

As a follow-up: not really going to push the whole is-satellite/kite imagery maps point. You can email Google Maps and have them rename the satellite maps to remote sensing data.

As far as the OSM/Yahoo connection: yes, it isn't absolute. But the reality is that a huge amount of new data is being generated from satellite footage - the new contributors are becoming couch junkies with free time instead of people with GPS strapped to their bikes. Cite as evidence the fact that much of the Haiti mapping effort was driven by sat. data either donated by GeoEye (via Google), or donated from the World Bank. This is awesome and it doesn't change the fact that great data is coming out of these areas, but it would also be awesome if we had a more people-based way of collecting this sort of data - and this is what kite mapping aims to do. It's not for the world, and it's not vector maps, but it collects data from vital areas and can connect that to the couch junkies with the time to trace it.

I think it's kind of absurd to think that this is going to duplicate the efforts of OpenStreetMap. It is being traced into OpenStreetMap (cartagen provides OSM as a base layer), and the developers are insanely aware of OSM - in fact a bleeding-edge user is on the OSM foundation. This isn't NIH syndrome... OpenStreetMap is doing an entirely different thing. Kite mapping = raster data from which one can contribute to OpenStreetMap. OpenStreetMap = vector data which can be traced from kite maps.
posted by tmcw at 7:26 AM on April 24, 2010

« Older Longtime voice of GEICO commercials fired after...   |   Unsung Heroine Newer »

This thread has been archived and is closed to new comments