Description of mask csv files
The masks are originally output as csv files, which are simple
ASCII files with comma-separated columns. The mask csv files
are then loaded into the Catalog Archive Server, which
will be the best means of accessing and using the imaging
masks. However, here we also give a brief description of the
mask csv files, as they are also available to download and use.
The mask csv's are available from the web, under the
DR5 root imaging directories for the
reruns. The csv files are located in the "stripe*/mask" subdirectories,
and are named "mask-$run-$camcol.csv", i.e. one file for each of the
6 camera columns of each imaging run.
The format of the csv files is detailed on the SDSS data model page,
in the entry for
As mentioned before, the masks are in the form of convex polygons, and
the RA and DEC coordinates of the vertices of each mask are given
in the "area" column, which consists of the list of vertex coordinates
enclosed in quotes. Alternatively, each mask may also be more roughly
approximated by the bounding circle, whose center is given by the
"ra" and "dec" columns, and whose radius by the "radius" column.
The polygon masks have been truncated as needed to the PRIMARY survey area for
each survey field/filter (but note the bounding circles have not been).
So for a given filter, the masks will not overlap between different
survey fields. This does not hold true for masks from different filters,
because of details in the filter-dependent astrometric solutions.
Moreover, note that within a given field and filter, masks of the
same type should not overlap each other, but masks of different types
may. Hence one may need to resolve such overlaps when combining
masks of different types (such as BRIGHT_STAR and BLEEDING), for example,
if one wants to compute the total masked area.
For those who are unfamiliar with convex polygons and their
related computational algorithms, one resource is
J. Gosper's M.Sc. thesis . In particular, the description
2D convex hull algorithms may be useful
for merging overlapping masks (i.e. by finding the convex hull which
includes all the masks to be merged). Also, the section on
provides descriptions of useful algorithms for finding the area of a convex
polygon and for determining whether a given point lies within a particular
Details about the mask creation can be found in the algorithms page on masks.
Last modified: Wed Jun 21 18:40:21 BST 2006