Rome in a Day Project: 3D reconstruction from Flickr-found photos
A group of technology researchers in Washington state are attempting to construct a three-dimensional model of Rome from photographs found on Flickr -- in one day.
Entering the search term Rome on Flickr returns more than two million photographs. This collection represents an increasingly complete photographic record of the city, capturing every popular site, facade, interior, fountain, sculpture, painting, cafe, and so forth. It also offers us an unprecedented opportunity to richly capture, explore and study the three dimensional shape of the city.Rome in a Day (via Dean Putney)In this project, we consider the problem of reconstructing entire cities from images harvested from the web. Our aim is to build a parallel distributed system that downloads all the images associated with a city, say Rome, from Flickr.com. After downloading, it matches these images to find common points and uses this information to compute the three dimensional structure of the city and the pose of the cameras that captured these images. All this to be done in a day.
the latest
latest episodes
Very cool idea and the preliminary results are already impressive. Although they might get even more images if they search for "St. Peter's Basilica" instead of "St. Peter's Cathedral."
Well that didn't show anything. What photographs?
@ 2
I think all those up-ended pyramids are the locations of all the cameras that took all the tiny photos being combined together.
http://www.digitalartform.com/archives/2004/11/nodal_point_pan_2.html
This is awesome.
would be cooler if they didn't restrict themselves to just one day
If those camera views were all frames from a single moving camera then this would be what programs like Boujou and Syntheyes and PF Track do.
They all generate an object/environment point cloud as part of their camera tracking process.
I am somewhat surprised no one has mentioned Microsoft's Photosynth: http://photosynth.net/Search.aspx?keyword=rome&sortby=Best%20Synth
BoingBoing previously covered this technology in a demonstration at TED (complete with Rome flickr photos): http://www.boingboing.net/2008/06/26/top-10-ted-talks.html
The automated aspect of it is related to this -
http://www.graficaobscura.com/merge/index.html
Interesting ... but more interesting (and practical) to me is the work of another group, "Finding Paths through the World's Photos":
http://phototour.cs.washington.edu/findingpaths/
creating a spacio- view (that could also be temporal) using multi-person exposures of 3D objects.
The possibilities for virtual tourism are wide open ... most of us can't haul ourselves everywhere we'd like to. Mob annotations would be great for education. The addition of ambient sound would be even better ... if more people would only make recordings of spaces!!
The innovation here isn't what's being done it's HOW it's being done:
"The key contributions of our work is a new, parallel distributed matching system that can match massive collections of images very quickly "
I assume the presentation layer is similar to Photosynth (they don't say, but it looks exactly like it). As such, this might fill in the gab that the demonstrations of Photosynth several years ago glossed over: that it took yonks to knit the pretty pictures together.
PS: Anyone who wants to see more of the pretty bit, as opposed to the data-crunching part that this research is about, should follow the links provided by #7.
@7 - The first thing I thought of when I saw the title was Photosynth. I haven't heard about it for a while, but the other software demo'd at the TED that year, Seadragon, cropped up on iPhone. That stuff was slick. I guess it's just waiting for the killer app to come along and make it useful.
For instance, PDF viewers are complete dogmeat. If I could point Seadragon at a directory structure full of PDFs and images, docs, whatever and flip through them and zoom in the same, that'd would be the killer app. Why isn't it baked into Windows 7? All this innovation just flapping in the wind.