Computer scientists from the Georgia Institute of Technology and Microsoft Research have developed 4D Cities, a software that shows the evolution of a city over time. New Scientist writes that you can see a city change in four dimensions. So far, the team has only modeled Downtown Atlanta by scanning historical photos. The software automatically sorts these snapshots into date order and then builds an animated 3D model that shows changes. This application will be useful for architects, historians or town planners. Now the researchers want to develop models for other cities.
The team is currently working on "a small 4D-pilot project centered around the Candler building in Downtown Atlanta, which is the trapezoid-shaped building at the center of the map above, using Microsoft's Virtual Earth API" (Credit: Georgia Tech). An interactive version of this map is available from the 4D Cities website by clicking on the "4D-Atlanta" link. And here is a link to a larger, standalone map where you can choose between 3 kinds of interactive views (road, aerial and bird's eye) -- and have lots of fun (especially if you know Atlanta).
The software has been developed at Georgia Tech by Frank Dellaert and Grant Schindler, with the help of Sing Bing Kang of Microsoft Research.
Here is how they introduce the project on the 4D Cities home page. "The research described here aims at building time-varying 3D models that can serve to pull together large collections of images pertaining to the appearance, evolution, and events surrounding one place or artifact over time, as exemplified by the 4D Cities project: the completely automatic construction of a 4D database showing the evolution over time of a single city."
Now, let's return to the New Scientist article for additional details.
To create a model of Atlanta, the researchers scanned in numerous historical photos of the city that had been snapped from similar vantage points. The software is designed to identify the 3D structures within the image and break them down into a series of points. It then compares the view in each one to work out why some of these points are visible in some of the images but not others. Was the building simply out of shot? Or was the view of one building blocked by another? "If we can rule out those two possibilities, then we know that the reason we don't see a building is because it didn't exist when the image was taken. Either it was not yet built or it had already been demolished," says Schindler.
For more information about this project, you should read the talk that Frank Dellaert gave at the 3rd International Symposium on 3D Data Processing, Visualization & Transmission 2006 (3DPVT 2006) (PDF format, 30 pages, 8.32 MB). You also should read a paper named "Inferring Temporal Order of Images From 3D Structure" (PDF format, 7 pages, 1.78 MB), which was presented at the IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2007).
Sources: Paul Marks, New Scientist Magazine issue 2610, June 27, 2007; and various websites
You'll find related stories by following the links below.
1:57:59 PM
Permalink
|
|
