Feature is basically tie point and the same can be visible from more cameras (and basically needs to be visible).
It is possible, you just need to know the pixel size and provide some corrections to the equation.
Image name is basically a camera name for this case.