Given; (1) We have a stationary camera at an unknown position (Cx,Cy,Cz) and unknown attitude ie.angles (Cxa, Cya, C,za) looking at a scene. (Imagine camera on top of a high building -we don't know the building coordinate- looking at the town) (2) Camera resolution is NxM (800x600) (3) In the camera view we can identify a set of Objects {Oi: i=1,...K,
(A) You should search on "camera calibration" and photogrammetry. (B) Given (A) you can simply map the bounds of the image plane to world vectors (C) Although you have a mapping O --> P you lack data to fully map P --> O because you lack depth data in the 2D image. The mapping gives you a vector of unknown length. If you can define a ground plane the intersection with the vector gives you a possible 3D location for the object. You could also look at stereography and other 3D imaging techniques.
Graham
Field of view is based on lens geometry. If you know the focal length, then it's easy. One example web site:
Do your own homework! :)
Have something to add? Share your thoughts — no account required.
Ask the community — no account required