Identify clumps within a 3D FITS datacube.
Direct download link (GitHub): quickclump-master.zip
To find clumps in your FITS datacube, type
$ python qc.py my_datacube.fits
To show the usage help, type
$ python qc.py -h
To run Quickclump in interactive mode, type
$ python >>> import qc >>> qc.main(["my_datacube.fits"])
Quickclump is an automated, fast and accurate clump-finding code. It was
inspired by another clump-finding code called
DENDROFIND (Wunsch et al.,
reaches the same accuracy as DENDROFIND run with the parameter
Nlevels set to
infinity. It also runs much faster and its running time and memory usage scale
linearly with the input datacube size.
Quickclump was originally written for analysing astronomical data -- radio observations of interstellar clouds -- but its use is not limited to astronomy. Quickclump doesn't presume anything about the origin of the input data. It will find clumps in any 3-dimensional rectangular array served in a FITS file.
Running Quickclump on the whole LAB HI Survey (721x361x485 pixels)
(Kalberla et al., 2005)
on CPU Intel i7-4770S took 10 minutes and the memory utilisation peaked at
870 MiB. It found 4178 clumps. Quickclump was run with option
--otext None, i.e. producing no text output.
Tests with real CO data from the
Galactic Ring Survey
(Jackson et al., 2006)
show that Quickclump consumes up to 10 times the size of the input datacube.
std() method is especially eager for memory and takes about 6 times
the size of the array (input datacube). However, if you provide the
--Tcutoff at the command-line, the memory-hungry
numpy routines won't be called and the memory usage should stay below 5 times
the size of your input datacube.