dstndstn / astrometry.net

Astrometry.net -- automatic recognition of astronomical images
http://astrometry.net
Other
682 stars 189 forks source link

Orientation reported incorrectly -- solved? #143

Open mcovington opened 6 years ago

mcovington commented 6 years ago

This may be a solved problem, as it seems to have gone away from the online server version, but is still present when I run the Debian distributed versions (0.67 and 0.70).

Picture orientation is often or always reported in mirror-imaged form, e.g., "Up is 175 degrees from east" when in fact it is 5 degrees from east. At first I thought this was a simple 180-degree error because I was working with images that were nearly straight north-south.

More here. Read down the thread to see more about it, because the first few messages do not characterize it accurately. https://www.cloudynights.com/topic/638573-astrometrynet-orientation-180-degrees-off/

mcovington commented 6 years ago

Related to #116 I see after doing some digging. I am assuming the fix simply hasn't propagated into the Debian package yet.

dstndstn commented 6 years ago

That fix was only for what the website shows, not for the command-line version.

dstndstn commented 6 years ago

It's almost certainly the issue that we convert JPEG and other image formats to FITS, and there is often a vertical flip involved in this process. The reported orientation is for the FITS file, thus the error.

Saturnyn commented 5 years ago

@dstndstn Hi, sorry for intruding in the conversation, can you confirm that the vertical flip will always happen whenever a jpeg file is used ? (In that case a simple fix for me will be to flip back the orientation output by astrometry.net.)

As a feature request: I think it would be really nice if you could find a way to fix the problem directly inside astrometry.net rather than at the website level. At the moment, having a different orientation on the two sides makes things a bit confusing.

In any case, thanks for the great software, it is really impressive to see it output such precise results from the worst images we throw at it.