This is a photo that I took in Wales in March, on the Pembrokeshire coastal path.
The camera (actually a Samsung phone) was only about 20 - 30 metres above sea level when this was taken. But I wondered if I'd be able to see the curve of the horizon if I compressed the pic sideways, like this (image is cropped but shows the whole breadth of the first image:
Although it's tempting to blame the curvature present in the image on lens distortion, I don't think that's the case here. The horizon was actually slightly below the centre of the image, so any barrel distortion would tend to suppress the curve, not exaggerate it. However I have no idea how to do the maths for the degree of curvature you'd expect to see from that elevation.
However, I suspect that is just an exaggerated view of the Earth's curvature.
Anyone any thoughts?
From the Samsung SM-A528B user manual:
Your device’s camera features a wide-angle lens. Minor distortion may occur in wide-angle pictures or videos and does not indicate device performance problems. To correct the distortion in pictures, tap ⚙ on the preview screen, tap Format and advanced options, and then tap the Ultra wide shape correction switch to activate it.
Did you...?
You do realize that you have a decent percentage of TZ-UKers holding rulers to their screens?
My result was that in the larger picture, the centre of the horizon is slightly higher than at the sides; presumably this does demonstrate that the earth isn't flat.
I didn't. But the wide angle distortion described there would act against the curve visible in the image, it wouldn't enhance it.Originally Posted by PickleB
See this article, please, if seeing really is believing, here’s a few steps you can take to see the curve and in the comments: "Mainstream science tells us we can’t see the curvature until somewhere around 60,000-130,000 feet."
The very existence of the horizon is an illustration of the Earth's, curvature (perpendicular to the one your compression appears to show); should you climb to a pole or mast, you would see further (hence crows nests in sailing ships).
'Against stupidity, the gods themselves struggle in vain' - Schiller.
Why do you say "the wide angle distortion described there would act against the curve visible in the image, it wouldn't enhance it", please? I ask as I cannot find any description of the distortion in that user manual. Have you heard of barrel distortion...see this WikiLink?
Again, from the user manual, "Customising camera setting...Pictures...Ultra wide shape correction: Set the device to correct distortion in pictures taken with the Ultra wide camera." Again...did you?
The Earth is round?
I believe he says that because the horizon is in the lower half of the picture, when the expected distortion would curve the horizon the other way. I am not sure that assumption is correct even if it is a reasonable one to make.
'Against stupidity, the gods themselves struggle in vain' - Schiller.
Pembrokeshire is beautiful
Sent from my M2101K6G using Tapatalk
Again: no, I didn't.
So: here's a graphic that represents what barrel distortion does to an image:
In this case, the line of the horizon - because it's below the centre of the image - is analogous to one of the lines below the centre of the graphic on the left. But the horizon line in the compressed image I posted curves upwards. Hence, if barrel distortion has compromised the image at all, then it has suppressed the curve. It hasn't caused it, or exaggerated it.
I'm open to the possibility that the second image has been influenced by pincushion distortion, but I don't see how a relatively wide lens on a phone would do that.
In which case, if the photo is uncropped, I should have referred the OP to pincushion distortion rather than barrel.
The best way to investigate the presence (or absence) of distortion would be if the OP were to return to the Pembrokeshire Coast National Park and take a series of photo with the horizon at different heights in the frame.
If you look at the StreetView photos associated with the link above and scan left to right, you can see that the horizon in not consistently flat or, apparently, slightly curved. To me, that indicates that Google are aware of lens distortion and apply some correction selectively.
I disagree the earth is flat and I have a picture to prove it!
“ Ford... you're turning into a penguin. Stop it.” HHGTTG
No. It's an oblate spheriod. Balanced on the back back of turtles
Sent from my SM-G991B using Tapatalk
You can't just say the earth isn't flat!.
I need some facts to prove this please.
thats the lens distorting, even very fancy huge pro lenses do it and it is corrected in the camera or software
No no, it's merely curved
Sent from my DN2103 using Tapatalk
However, I suspect that is just an exaggerated view of the Earth's curvature.
Imagine a this protractor is the size of the earth, the distance between one degree is 111km. I'm not sure of the distance between the far left and right of your image, but at 20m high the horizon would be around 16km away, so perhaps 32km between the edges of your image wouldn't be enough to see a curve.
If the Earth really was flat cats would have pushed everything off it by now.I disagree the earth is flat and I have a picture to prove it!
Started out with nothing. Still have most of it left.
If that is Pembrokeshire then what you are seeing is the Hump Back Wales.
"Bite my shiny metal ass."
- Bender Bending Rodríguez
I'm trying not to, but: wide angle lenses distort in a particular way that wouldn't cause the upward curvature of a line below the centre of the image. I'm sorry if you haven't understood that.
Yes, this is well understood. I wonder if the curvature of the horizon in the image is a product of over-correction by the phone's camera app? That at least would make sense.
I think an excellent experiment would be to take a pic on a clear day with a better camera, with the horizon as centred as possible. Then to do the same again, with the camera upside down. Then to compare the two images. Next time I have a clear view of a horizon from a bit of elevation, I'll do that.
Posting Permissions
Reply With Quote
