Human Viewable (color composite) Image creation.

Creates a composite colour image from sets of input FITS files, following the Magain, Courbin & Sohy (1998) deconvolution algorithm and the Lupton et al (2004) composition algorithm (with extensions by Hogg & Wherry.)

Phil Marshall (Oxford) [email protected] Cato Sandford (NYU) Anupreeta More (IPMU)

October, 2012

There are two top level scripts. deconvolve.py takes in some number of FITS files (or a directory containing FITS files) and deconvolves each image to a Gaussian PSF. compose.py takes in 3 FITS files as input, and returns a color composite, color-saturated png image with an arcsinh stretch.

I have an image in three bandpasses, called i.fits, r.fits and g.fits, in the current directory. I wish to deconvolve each such that the PSF is a circularly-symmetrical Gaussian with \sigma=2 pixels.

python deconvolve.py -p 2 i.fits r.fits g.fits

Now I wish to combine the three new images.

python compose.py gDeconvolved_i.fits gDeconvolved_r.fits gDeconvolved_g.fits

View the result:

display output.png

The composition script requires:

• the Python Image Library (PIL) available from http://www.pythonware.com/products/pil/

The dconvolution script requires:

Both require:

• numpy