The phenomenal growth of photo sharing websites like Flickr and Picasa has spawned a new art. With more than 80 million photos being uploaded to photo sharing websites every day, computer scientists have invented a technique that automatically creates 3-D models of famous landmarks and geographical locations, using ordinary two-dimensional pictures captured by tourists and enthusiasts.Thus Rome has been built in a single day, on a PC, using a method of image clustering, stereo, stereo fusion and structure from motion to give high computational performance.
It was devised by a team of researchers from the University of North Carolina at Chapel Hill and the Swiss university, ETH-Zurich, who believe the system will help preserve heritage sites, by allowing tourists to explore sites virtually.
Led by Jan-Michael Frahm, research assistant professor of computer science in the UNC College of Arts and Sciences, 3 million images of Rome available online were used to reconstruct all of the city’s major landmarks.
Although the technique of using millions of photos to recreate detailed scenes has been demonstrated by other systems such as Microsoft’s PhotoSynth, Frahm says they required clusters of computers to perform the necessary image analysis. Frahm said his team used a simple home PC equipped with four powerful graphics cards, giving an improvement of more than a factor of 1,000 over current commercial systems.
“Our technique would be the equivalent of processing a stack of photos as high as the 828-meter Dubai Towers, using a single PC, versus the next best technique, which is the equivalent of processing a stack of photos 42 meters tall – as high as the ceiling of Notre Dame – using 62 PCs,” he said. “This efficiency is essential if one is to fully utilise the billions of user-provided images continuously being uploaded to the Internet.”
One of the advantages of the 3-D models compared to viewing a video of a landmark is that the Internet photo collections used to construct them show the scene at different times and under different lighting and weather conditions, creating a richer experience for viewers, said Frahm. Video can be used as well, and usually shortens the processing time needed for reconstructing the models, he said.
Frahm said the models could eventually be embedded into common consumer applications such as Google Earth or Bing Maps, allowing users to explore cities from the comfort of their homes. Such applications could prove useful to travellers.
“You might be able to take a picture with your cell phone of a monument that would not only give you information about that monument, identifying it from the image, but could also tell you your location more precisely than even GPS,” Frahm said.
Frahm said the technology could also be a building block for disaster response software. For example, an aircraft could be sent to take video of the aftermath of a hurricane, and the resulting 3-D model could be used to assess damage from a remote location, saving time and money.