Input image camera position

[JennySeidenschwarz]

Hey :)

the input image camera position should actually be the transformation matrix we get with elevation=0 and azimuth=0, right? Additionally, if we do have the distance of the camera to the object, we could add this distance to the transformation matrix generation, right? If we then remove the scaling of the input image, in theory the scale of the mesh should be the "correct" scale.

Best,
Jenny