You can get a good sense of the grandeur of Rome's Colosseum by looking at photos, but disparate snapshots give you a real idea of the ancient ruins scattered throughout Italy's Eternal City? They can now, thanks to a team from the University of Washington in Seattle, that has developed a method of creating accurate 3D simulations of entire cities using thousands of photos.
By submitting your email address, you agree to receive emails regarding relevant topic offers from TechTarget and its partners. You can withdraw your consent at any time. Contact TechTarget at 275 Grove Street, Newton, MA.
Sameer Agarwal's team was able to create a simulation of Rome using 150,000 images harvested from photo-sharing website Flickr, and build a virtual model within a day.
A simulation of Rome was built from 150,000 images harvested from photo-sharing website Flickr
The principle isn't new - the team previously developed tools which can create 3D models from a collection of photos, which subsequently evolved into Microsoft's Photosynth. But while that technology was good at using snapshots of single tourist attractions, it was unsuited to tackling larger projects, such as recreating cities. "Using the existing system it would have taken years to recreate a whole city," says Agarwal.
The team has developed methods to improve the process of matching images and performed the 3D modelling using a cluster of about 500 computers working in parallel. They also created models of Venice using 250,000 snaps and Dubrovnik using just 60,000. They hope the work will help us preserve and remember the beauty of today's cities for future generations to enjoy.