next previous
Up: The ESO-Sculptor faint

5. Astronomical coordinates

To obtain precise astrometry for the detected objects in our CCD frames, the regions of the SERC-J photographic plates containing our survey were digitized at the MAMA ("Machine Automatique à Mesurer pour l'Astronomie'' at Paris Observatory) which is developed and operated by CNRS/INSU (Institut National des Sciences de l'Univers).

The astrometry is performed in two steps.
(1) By alignment of the images of the digitized plates onto our CCD frames (using existing MIDAS commands applied to objects in the field), we are able to derive the coordinate transformation equations from CCD to photographic plate. There is one set of transformation equation for each CCD frame.
(2) The transformation equations from photographic plate to equatorial coordinates are provided by the MAMA facility as FORTRAN programs. The astrometric reduction uses the PPM catalogue (Roeser & Bastian 1991).
(3) Combination of the two transformations yields for each CCD frame the transformation equations from pixels to equatorial coordinates. We use a set of MIDAS procedures which were written with the goal to perform the transformation routinely for the numerous fields of the programme (Revenu & de Lapparent 1992).
The overlapping regions of neighbouring CCD frames allow us to estimate the internal astrometric consistency. In Fig. 4 (click here), we plot the median position difference defined as tex2html_wrap_inline2609 in bin sizes of 1 mag. The NTT observations are represented by dots and those from the 3.60 m by crosses. The error bars are the 1 tex2html_wrap_inline2611 dispersion measured in each magnitude bin. For the NTT, the differences of median position vary from tex2html_wrap_inline2613 for tex2html_wrap_inline2615 to tex2html_wrap_inline2617 for tex2html_wrap_inline2619, and from tex2html_wrap_inline2621 for tex2html_wrap_inline2623 to tex2html_wrap_inline2625 for tex2html_wrap_inline2627 with the 3.60 m. The better accuracy for the NTT fields is due to the better sampling (smaller pixel size) which improves the measurement of the central position of the objects: the seeing disk is spread over tex2html_wrap_inline2629 pixels for NTT frames, whereas it is spread over only tex2html_wrap_inline2631 pixels for 3.60 m frames.

  figure386
Figure 4: Mean difference in arcsec in the astrometric position of common objects in CCD overlaps as a function of R magnitude in bin sizes of 1 mag. The data for the NTT and 3.60 m telescopes are represented by dots and crosses respectively. The error bars show the 1 tex2html_wrap_inline2635 rms dispersion


next previous
Up: The ESO-Sculptor faint

Copyright by the European Southern Observatory (ESO)
web@ed-phys.fr