Re: [OT] Mathematical photographers that program (Perl)? (Updated.)

[Your Mother]

FWIW, the image looks CG to me. I am not asserting it is, just offering that as a possibility if it might matter.

Edge detection of straight lines (edges of bolt heads) could give you the horizon line at least (assumptions being they and the plane they rest on are regular/level).

[BrowserUk]

FWIW, the image looks CG to me.

Even with all the nicks and scratches?

My thoughts are that they've probably been shot-peened or sand tumbled.

I've added an image I've just taken myself to the root node.

[Your Mother]

Yes; I agree about the tumbling but artificial flaws are a good trick to fool on expectations. It’s the quality of the light on the metal that looks off to me. I’m not at all sure, though. It might be just a filter/tart-up edit on it. We’re only another few years away from never being able to know again if still images, and then video and audio, are authentic. o_O

[Anonymous Monk]

There's a point midway along the bolt where the threads look deepest and straightest. However, that's not "the central axis of the lens," it's the point where the axis of the bolt is perpendicular to a line drawn to the focal point of the lens. Try this: take a picture of two similar bolts, but put a few sheets of paper under the corner of one, so it leans toward (or away from) the camera a bit. You'll see that the "deepest and straightest" point is displaced upward (or downward).

[BrowserUk]

There's a point midway along the bolt where the threads look deepest and straightest.

Yes. That really stands out with the short focal length lens of a phone camera doesn't it. It's not so obvious with the professional shots.