I'm beginning some research in 3D position finding using cameras. I would like to try the technique where I take a picture of a known 2D symbol and work out camera position from the size, scale and skew of that symbol. I know this technique is used in Augmented Reality to determine where the camera is and where to place the visual augmentations. I have no experience in AR software, so I don't know what to call this technique (to search for info on it) and don't know if there are open source libraries I can use for getting camera location with this technique.
Alternately, are there any computer vision algorithms that can locate a repeatable location on a symbol so that I could identify 3+ points and do the math myself?
- What is the terminology for these techniques that I can search for?
- Are there existing open source packages that do this?
(If it matters, the symbols I would be using are irregularly shaped (not clean rectangles or squares like a QR code) and there will be plenty of unknown visual clutter around the symbol. Imagine the symbol on a scrap of paper that is on top of a graffiti-covered wall. The symbol will contrast from its immediate surroundings.)