The RA and DEC information was converted back to X- and Y- coordinates using the
same ZPN projection and a scatter plot was produced for each WFC pointing
with the CCD outlines overlaid (CCD bounadaries were deduced from the FITS
headers). This allowed us both to detect cases where, for instance, a very
bright object had been removed from an image leaving a blank area
(see Fig. 
(b)), but also to
check the conversion from RA and DEC to ZPN coordinates we were performing, as
it would be of importance later on.
Figure: i--band spatial distribution plots for the SDS-1 field, showing (a) all objects, (b) stars only, (c) galaxies only and (d) objects classified as noise only.
Figure: Example of imperfections shown by the spatial distribution plots: (a)z--band G 133 field (galaxies only) showing fringes, (b)z-band, MKN 205 field (all objects) showing the effect of removing a bright object, (c)i--band, GRB 001025 field (all objects) showing a very crowded low galactic latitude field, (d)g--band, EQ PEG field (galaxies only) showing a small defect in the
CCD on the left (also seen in (b)).
While the Fig. is a fairly typical example of the
distribution of objects over the image and across different object classes
examples of some of the defects detected are shown on Fig. .
Fig. (a) shows very strong fringe patterns, which are seen on
an important fraction of the images (30%), and usually classified as
galaxies. Note that the defringing is still under way
(see [3]).
Fig. (b) shows a case where a bright object was removed.
Fig. (c) shows a low galactic field which is very crowded.
Fig. (c) shows a small defect on one of the CCDs.