CUPID - A 3D Clump Identification and Analysis Package

What is CUPID:

CUPID is a package of commands that allows the identification and analysis of clumps of emission within 1, 2 or 3 dimensional data arrays. It is being developed by the Joint Astronomy Centre (JAC), Hawaii, as part of the SCUBA-2 and HARP advanced data products survey pipelines. Whilst targeted primarily at sub-mm cubes, it can be used on any regularly gridded 1, 2 or 3D data.

A variety of clump finding algorithms are implemented within CUPID, including the established ClumpFind and GaussClumps (see Stutzki & Guesten 1990, ApJ 356, 513.) algorithms. In addition, two new algorithms called FellWalker and Reinhold are also provided. New algorithms can be added as needed.

One of the goals of CUPID is to allow easy inter-comparison between the results of different algorithms. For this reason, the catalogues produced by each algorithm contains a standard set of columns containing clump peak position, clump centroid position, the integrated data value within the clump, clump volume, and the dimensions of the clump. In addition, pixel masks are produced identifying which input pixels contribute to each clump.

At the moment, the main functionality of the package is in the areas of background estimation and clump identification. CUPID integrates with the Starlink Software Collection, and so the facilities of (say) the KAPPA package can be used to visualise the gridded results, whilst TOPCAT can be used to explore the catalogues. The native data format used by CUPID is the Starlink NDF, but FITS data can be handled transparently by means of the Starlink CONVERT package.

Obtaining CUPID:

Pre-built binaries for CUPID are available as part of the JACH release of the Starlink Software Collection. The latest cutting-edge development build can be obtained from http://starlink.jach.hawaii.edu/starlink/rsyncStarlink. The source code can be browsed as part of the Starlink GIT repository.

Using CUPID:

If you install the Starlink software collection to, say, /home/fred/starlink, you can initialise the CUPID package by doing:

% setenv STARLINK_DIR /home/fred/starlink
% source $STARLINK_DIR/login
% source $STARLINK_DIR/cshrc
% cupid

% export STARLINK_DIR=/home/fred/starlink
% source $STARLINK_DIR/profile
% cupid

CUPID commands can then be entered at the shell prompt, following by an optional list of "parameter=value" strings (prompts will be given for any required parameters that are not supplied on the command line). For instance:

% findclumps in=mydata out=clump-mask method=fellwalker

Documentation:

More detailed information on the CUPID commands and how to use them is available in the draft Starlink User Note (SUN) 255.

Other Information:

  1. A very brief introduction to the "min-max" filter used by the CUPID:FINDBACK command.

  2. A very brief introduction to the FellWalker algorithm.

  3. A poster describing CUPID presented at ADASS 2006.

Starlink: CUPID (last edited 2014-07-16 09:48:45 by DavidBerry)