Ok all my followers check out the coolest thing since sliced potato's. A group of computer nerds (no offense) have created a 3D model of Rome using only pictures pulled off of Google. Here the video:
Wow, it blew my mind the first time I watched it too. For more a more finished project and cool Google Earth layout check out the very cool website HERE. Now if you don't understand why this is so cool let me try and break it down for you. Pictures are essentially a huge grid of cells with numbers in them, each cell corresponding to pixel in your computer screen. The number in the cell tells it what color the pixel should display. Put a bunch of those together and boom we get a picture. Simple. Thats the super simplified version but it will do.
Knowing all that, we know that pictures are a 2D representation of what we took a picture of with do information about depth. Given any picture we cannot find out how far say a tree is in that picture from your camera. So how do these guys create 3D models like these with just pictures? Well its complicated, but a quick answer would be with lots of images. Using bunches of them they can find out through transformation matrices etc. etc. where the cameras took each picture and slowly piece together a kind of point model representation. (watch the video if you dont understand) After they got a kind of skeleton made of thousands of points they lay bits of image over them like a skin. The result is you can see in 3D what a statue or building looks like from 6000 miles away. Who would have thought? I love this kind of innovation and have to tip my hat to the UNC guys for this amazing project. Hope you enjoyed it as much as I did - Laters
No comments:
Post a Comment