upload problem

[Guang Yang]

Hi,

  When I attempted to upload my fits file (url mode), the page told "An internal server error occurred. Please try again later."

  The url is http://www.asiaa.sinica.edu.tw/~whwang/goodsn_ks/DATA1/HDF_K_uJy.fits.gz

  Is there any other way to upload my image?  

  PS: I'm in China, and direct file uploading is too slow.

 Thanks.

 Guang

[Dustin Lang]

Thanks for your bug report!  It turned out that we were trying to retrieve the given URL immediately, but since it took a while, the web server proxy would time out waiting.

I fixed that in our development version (supernova.astrometry.net) and submitted your image; it worked, and the solution is available here:
  http://supernova.astrometry.net/user_images/6734

That fix will be available in nova.astrometry.net the next time I do a launch.

cheers,
dustin

[Guang Yang]

Hi, Dustin

Thanks for fixing the bug. :) 

Would you mind answering me another question?

Now, I'm doing a project about one specific region, GOODS-N, and I have collected many images. Some of them can be downloaded from others' server.

However, others are mosaiced by myself, so I can't upload them via url (I don't have a big server -.- ). Is there any solution for me to upload those giant files to astrometry.net?   

Good luck,

Guang

  

[Dustin Lang]

Hi,

You could run a source detection routine locally and then upload just the list of detected source positions.

Say, run SExtractor or the "image2xy" program in the Astrometry.net code and upload the FITS table of x,y positions.

cheers,
--dustin

[Guang Yang]

Dear Dustin,

OK, I will try that. :)

regards, 

guang.

[Guang Yang]

Hi, Dustin

  

I'm sorry to bother you again, but another problem has arisen.

After uploading the package (a zip file including S.E. position results and img headers), it tells 'Unfortunately, we couldn't process this submission properly'

see http://supernova.astrometry.net/status/6761

good luck,

guang

[Dustin Lang]

Could you please try uploading just a single file -- a .cat file, no need for the hdr.fits files.

Two issues though --

1. We need to have some way of sorting the stars by brightness.  You must include at least one SourceExtractor flux measurement.  I usually choose FLUX_AUTO.

2. Right now I think the web service is not smart about looking for a "flux" column.  I think it expects the column to be named "FLUX" (not "FLUX_AUTO").  So you will either have to rename the column in the FITS table  (maybe using "fitscopy" to create a new table with the FLUX_AUTO column renamed to FLUX, or "modhead" to change the name  of the FLUX_AUTO column).  OR, you can sort the files yourself before submitting them, so that the brightest sources are listed first.

cheers,
dustin

[Guang Yang]

Hi, Dustin:

Sorry, I didn't know flux is needed, as the uploading page says "FITS binary table, containing a BINTABLE of detected objects, with X and Y pixel positions in "D" (double) or "E" (float) columns, with one object per row". 

But the error message appeared again, after uploading a .cat file with 3 columns FLUX, X_IMAGE, Y_IMAGE. http://nova.astrometry.net/status/62290

Maybe the format is wrong again. Would you please send me a model .cat file, so that I can follow its format?

good luck,

guang

[Dustin Lang]

Oh, you're right, I forgot about X and Y too:

Just like you renamed FLUX_AUTO to FLUX, please rename X_IMAGE to X and Y_IMAGE to Y.

Sorry about that.

--dustin

[Guang Yang]

Unfortunately, the job failed, though it had been working for several minutes. 

At first, I thought it might be due to the large number of objects in the catalog, so I adopted a more strict criterion for detection in S.E and uploaded again.

However, it still failed. :( http://nova.astrometry.net/status/62296 )

[Dustin Lang]

Hmm, it looks like maybe the web service is not paying attention to the FLUX column.

I tried sorting your file by FLUX (brightest first) and got this solution:

http://nova.astrometry.net/user_images/72916#redgreen

So I guess you need to sort your files before submitting them.

(There is a program "tabsort" in the Astrometry.net tools for doing this)

Sorry for the trouble!

--dustin

[Guang Yang]

COOL!

It's really really amazing. 

In fact, recently I'm correcting astrometry by myself, planing to make use of the old geomap and geotran in IRAF.

I have carefully matched my catalog with USNOB1.0 (ruling out saturated stars, large-proper-motion stars, etc), and roughly found some mean offset in ra,dec. 

Surprisingly, the results you have just delivered are quite the same with mine, and you have nothing but (x,y) and flux.

Previously, I thought that there is only linear transformation available in the fits header, as my e-mail told. (IRAF says it can only reflect linear transformation to the header)

Now, I tend to adopt astromety.net for astrometry calibration in my work, rather than do that by myself. 

Your work is indeed useful and I will introduce this great tools to my friends and collaborators.

thank you :)

regards,

guang

在 2013年8月28日星期三UTC+8下午2时35分23秒，Guang Yang写道：