Re: From a camera's point of view *LINK*
JKJ
>To figure out exactly what edge the camera will see, draw a circumference of the object (top view) on a large piece of paper. Place a point the same distance away as your camera, or more precisely, the nodal point of the lens. (Some cameras have the nodal point marked to aid with panorama photography.) Draw two straight lines from that point tangent to either side of the circumference. From this, you can see the error is higher with a wide angle lens since the camera is closer for the same image width.
Note that this shows the edge the lens will see and has nothing to do with any distortion in the lens, such as barrel or pincushion, which is a different issue entirely.
You can draw similar lines from the side view to see the top and bottom edges that will be captured.
Fortunately, the edge the camera captures is the same as you see with your eyes at that distance (or at least with one eye closed) so the image looks entirely natural except at extreme camera distances.
Dept. of Mostly Useless Information: There is a way to take the photo with a special camera that captures EXACTLY the edges on the diameter, 180 degrees apart. Using a vertical slit in a focal plane shutter curtain, the camera or object is moved in sync with the shutter such that light path through the slit is always perpendicular to the object. (50-60 years ago photographers used the slow FP curtain shutters in Speed Grafix 4x5 cameras to do tricks with moving objects!) This is somewhat similar to the way a flatbed scanner works with 2D objects and has been used with lasers for precision work with 3D objects.
Dept. of Remotely Related but Interesting Information: Forestry people use something called a Biltmore Stick to measure the diameter of a tree by sight. The markings on the stick compensate for the perspective error you mention. The diagaram under "Measuring Diameter" in the link below illustrates the angles, just as with a camera lens.
JKJ
Biltmore stick