How long does it take to create an accurate 3D model of a city on a computer? If you're familiar with virtual reality modeling, you know it can take weeks. But according to New Scientist, engineers from the University of California at Berkeley have found a speedier way to capture a city. Using a concept dubbed "virtualized reality," which mixes inputs from lasers to measure distances and digital cameras to scan a city, they can now build a 3D model of a whole city in about an hour. Obviously, the first applications will be military, but sooner or later, you'll be able to drive your car in an unknown city using this 3D technology. Read more...
Here is how New Scientist describes this concept of "virtualized reality."
The concept is similar to building a virtual reality model, but the process is very different. To produce a VR model, a programmer manually combines distance measurements and 2D pictures to make a 3D model. The new technique, dubbed "virtualised reality" by creator Avideh Zakhor, is automated and much faster.
"Right now, a detailed urban model can take many months to create," says Bruce Deal, vice-president of the Virginia engineering firm SET Associates, which is helping to adapt the technology for the US military.
"With the new model, we're talking about an hour or so." Virtualised reality scans the urban landscape using lasers and digital cameras mounted on a truck or plane. A laser measures distances to objects such as lamp posts and building facades, while the digital camera takes 2D photos. Another laser calculates the movement of the truck and checks its position against data collected from the aerial laser aboard the plane.
Below are two examples coming from the UC Berkeley Video and Image Processing Lab and in particular from this page about Fast 3D City Model Generation.
 |
Here is a first model of 3 downtown Berkeley blocks: the ground-based facade model has been fused with an airborne model (Credit: UC Berkeley Video and Image Processing Lab). |
|
 |
And this is another model of 3 downtown Berkeley blocks: the facade has been reconstructed and the foreground removed (Credit: UC Berkeley Video and Image Processing Lab). |
The above link contains other images and -- pretty huge -- VRML files you can download.
The research work has been published by several scientific journals. If you're interested in this subject, you should read "An Automated Method for Large-Scale, Ground-Based City Model Acquisition," published by the International Journal of Computer Vision (Vol. 60, No. 1, October 2004, pp. 5 - 24).
Here is a link to the abstract which starts like this.
In this paper, we describe an automated method for fast, ground-based acquisition of large-scale 3D city models. Our experimental set up consists of a truck equipped with one camera and two fast, inexpensive 2D laser scanners, being driven on city streets under normal traffic conditions. One scanner is mounted vertically to capture building facades, and the other one is mounted horizontally. Successive horizontal scans are matched with each other in order to determine an estimate of the vehicle's motion, and relative motion estimates are concatenated to form an initial path.
The scientists also add that it's pretty fast to scan a city like Berkeley.
A fairly accurate, textured 3D cof the downtown Berkeley area has been acquired in a matter of minutes, limited only by traffic conditions during the data acquisition phase. Subsequent automated processing time to accurately localize the acquisition vehicle is 235 minutes for a 37 minutes or 10.2 km drive, i.e. 23 minutes per kilometer.
And for even more information, here is a link to the full paper (PDF format, 20 pages, 1.4 MB).
Sources: Emily Singer, New Scientist Print Edition, May 5, 2005; and various websites
Related stories can be found in the following categories.
8:21:59 PM
Permalink
|
|
