M4 astrometry

[Antonio Kanaan]

hi All,

Not sure this forum is for help with unsolved fields, in case it is:

solve-field = Revision 0.76, date Mon_Sep_3_15:20:15_2018_-0400.

I am having trouble solving a field in M4.

 A friend of mine managed to get these fields to solve by tweaking the WCS in the original file.  If I work on his tweaked file ( https://drive.google.com/open?id=18rTwmnzHjXNAcW_J53SpunfOVqAm8XC_    -   luc-m4.HSCA17574604-sub1) I also get a solution.  If I work on the original file ( https://drive.google.com/open?id=1qHKZuh-rIBjK6smMl0mqlPOK0Y_5QOdt    -   HSCA16574604-sub1) no luck. 

I must be doing something wrong.  Astrometry.net does not need the image to be "almost solved" to work.  I have many times solved fields with no header at all, zero initial guess.  Which, by the way, I tried on these images feeding solve-field with the -y option to no avail.  

I also uploaded the images to astrometry.net, same results, my friend's modified image solves, the original does not.

Any hints appreciated, thanks

Antonio

[Dustin Lang]

Hi,

Could you please send a link to the version(s) you tried uploading to nova.astrometry.net

Which index files do you have?

Does the source extraction look okay?

It's not common for globular clusters to fail -- the reference catalogs tend to be pretty messy there.

cheers,

--dustin

--
You received this message because you are subscribed to the Google Groups "astrometry" group.
To unsubscribe from this group and stop receiving emails from it, send an email to astrometry+...@googlegroups.com.
Visit this group at https://groups.google.com/group/astrometry.
To view this discussion on the web visit https://groups.google.com/d/msgid/astrometry/e5333578-f1c9-4d40-818e-792e8e41a00c%40googlegroups.com.

[Antonio Kanaan]

hi Dustin,

Thanks for your prompt message. 

Yes, the source extraction seems fine on both cases.  

I guess you meant to say "not UNcommon" :-)

I am running on nova again with the exact same images I was running on my computer this morning.  The one with no solution is still running.

this is the success case, 

http://nova.astrometry.net/user_images/2870857#annotated

this is the failure,

http://nova.astrometry.net/user_images/2870869#original

cheers,

Antonio

[Dustin Lang]

Hi,

You need to set the scale range:

The default image size assumed is 0.1 to 180 degrees.  Your image is only 1.26 arcmin wide.  That is extremely narrow!!  I would not expect to find a solution, but with the default settings, it is *guaranteed* not to get a solution.  You would need to set the scale range (under "Advanced settings") to, say, a custom range of 1 to 1.5 arcmin.

But again, you shouldn't expect to get a result for images like this.  (Similarly, Astrometry.net doesn't typically work for Hubble Space Telescope images -- too small in angular size.) cheers, --dustin

[Antonio Kanaan]

thanks Dustin,

I tried playing with different scales, no luck.  It still puzzles me that astrometry solves the field fine with a better initial guess.  

I guess I will stop being anxious and wait till I do the data reduction and stitch up all different amplifiers + ccds, then I will have a huge field to run astrometry on.

thanks for your help 

Antonio

[Dustin Lang]

Hi,

When the image has an existing WCS, we try to *verify* it -- ask where we would expect to find stars and check whether stars are found there.

When the image does not have an existing WCS -- even if you give it a correct initial RA,Dec,radius estimate -- we need to search for matching quadrangles.  There is a minimum scale (2 arcmin) of quadrangles we can build (because reference catalogs don't have enough stars to go beyond that), so it would be quite unlikely to find a quadrangle inside your 1.3'x10' image.  That's why they don't solve.

cheers,

--dustin