You don't need to go that far. If you can get any two views of an object, you can get enough info to at least start to independantly establish the 3D shape. The closer these views are to being orthogonal, the fewer errors you will make, but cameras displaced by as little as 10-15 deg can give enough information for some shapes. Granted these would be more well understood or expected shapes, but it does work with a small parallax.Blappy wrote:Once have 3D cameras that can take a holographic image you might be able to have the printer make a part from that.
Extremely complex shapes need even more than two views, expecially if there are any concavities in the shape.
If you notice, in the little vidio, they used the method to make a model of the pipe image, but there was no hole in the pipe for tobacco! Same thing.
Paul