Print

Print
Dear Jean-Baptiste,

    There are two problems affecting the way gaia uses autoastrom.

    1. You are getting no objects back from CADC.
    The reason for this is that if you do NOT click "Initial WCS from Image", then gaia defaults to the CRPIX and CRVAL to derive the information about the location in the sky where to look for stars. Quite clearly, they are wrong by about 2 minutes in alpha. If you click "Initial WCS from image", then at least you get stars back from CADC.

    2. even if you get stars back from CADC, then there is another problem, which is that the RITmatch algorithm fails to find a suitable solution.
    In that case, what you have to do is:
    Enable the "advanced" mode, and then:
    Tuning -> Use match algorithm  -> NO (uncheck)
    Tuning -> No. Fit parameters -> 9
    Position -> Use WCS from image
  
    Doing this makes autoastrom in gaia find a good solution.

    Now, as to why the matching algorithm fails, that's a whole different story....

    Cheers,
    Luca



On 11/19/10 2:15 PM, Jean-Baptiste Marquette wrote:
[log in to unmask]" type="cite">Dear Gaia gurus,

I am experiencing a strange (and wrong) behaviour of the Automatic position matching function in Gaia.

I have a test image from the EROS-2 experiment which has been linearly calibrated using the Emmanuel Bertin's tool Scamp. So the header contains a classic CD matrix. The 2MASS catalog fits well on that image, with a slight distortion visible in the upper right corner.

I wanted to test AUTOASTROM with a non-linear fit. When I use 2MASS@CADC as reference catalog with the default WCS guess, I obtain no entry from that catalogue. So the process exits in error. I have to select Initial WCS from image to get 49 ref stars only, while the filed is very crowded. Plus the rms are high :

Best Fit Parameters
-------------------
nstars = 49            \ no. ref stars
rrms = 10.660375       \ radial rms, arcsec
xrms = 3.037770        \ x-axis rms, arcsec
yrms = 10.218392       \ y-axis, rms arcsec
plate = 0.535823       \ plate scale (mean), arcsec
prms = 19.895319       \ radial rms, pixels
nterms = 6             \ no. terms in fit
rarad = 1.392460038    \ projection pole Dec, radians
decrad = -1.135075897  \ projection pole RA, radians
rasex = 5:19:07.7      \ projection pole RA, sexagesimal
decsex = -65:02:06     \ projection pole Dec, sexagesimal

If I accept the solution, the 2MASS catalog does not fit anymore, and the plot symbols are badly distorted. If I switch to linear fit, it is wrong as well.

The image lm04510tbr9j1546.fits is available on ftp.iap.fr, directory pub/from_users/marquett

Thanks for your help,

================================================
Bien cordialement/Very truly yours/Mit freundlichen Gruessen,
Jean-Baptiste Marquette
Institut d'Astrophysique de Paris
CNRS - UMR 7095
Université Pierre & Marie Curie
98bis Bd Arago
75014 Paris - France
================================================