The CIAO roi tool
is used to convert say a source list into a stack of region
files, aka Region Of Interest (not return on investment).
The roi tool has a lot of options to control things like the
background region, including the detector edges via the field
of view file, etc.
When regions overlap it has several group options.
group=groupcreates 1 region file per bundle of sources that overlap.group=individualcreates 1 region file per source and ignores any overlaps.group=excludecreates 1 region file per source. Any overlapping sources are excluded from each other.
The exclude setting is the most common usages of roi. This is used
in scripts like srcflux. The complaint with this mode is that
any overlapping source area is never counted given how the region
logic is implemented.
The problem for roi is that it doesn't know how to choose which
overlapping source the overlapping area should go to.
FAP has determined in the context of CSC2, that the overlapping area should go to the brightest source.
Enter this script rank_roi. This script takes in a stack of roi
regions generated with group=exclude, assigns a metric to each
source region (in this case the number of counts; technically sum of
pixel values in input image), and modifies the roi files to assign any
overlapping area to the brightest source (highest metric).
The figure below shows a map ID'ing the source regions created using
roi (on the right), and with rank_roi applied (on the left).
The colors are random and map to the source ID; black colored pixels
unassigned to any source.
In the Right frame, the area between overlapping sources is unassigned (black). In the Left frame, the source regions have been ranked based on the number of counts and the area has been assigned to the brightest source.