A while ago I read about these guys who were able to write a very clever program that can regenerate a 3D model of a whole city from its photos on the web ...
The idea is that this program will go into a website like Flickr, download all the images tagged by the city name, and then attempt to deduce how the city looks like exactly, where each photo was taken, and rebuild the city in 3D!! The great thing about this is: a city can be built in one day, as the algorithm utilizes the power of parallel computing ...
When I first read this, I couldn't help but think about Google Earth. Google has attempted to introduce 3D into its Earth before. They have created sketchup, and wanted to utilize the community to create models building and monuments. You might also remember when Google modeled almost every building in Manhattan Island. The problem with this is that it has that human factor; building models are not always good, and are sometimes misplaced (I have seen a couple floating in the air).
This new technology promises to remove that human factor, but its currently limited to cities that have photographs covering the entire city (the reason why Rome and Venice where chosen). However, think of what the technology can do with the data from something like Google street view. This data is interesting because, not only it has excellent coverage, but also the exact location and direction of each photo is known.
Its thus good to see that Google was one of the supporters of this project :) ...
P.S.: I knew about this from: slashdot
No comments:
Post a Comment