ska_sdp_spectral_line_imaging.data_procs.deconvolution.radler module

ska_sdp_spectral_line_imaging.data_procs.deconvolution.radler.radler_deconvolve_channel(dirty_channel, psf, nx=None, ny=None, cell_size=None, algorithm='multiscale', niter=500, threshold=0.0001, gain=0.7, scales=None, **kwargs)[source]

Perform deconvolution using radler for a given channel data and returns restored model and the residual data

Parameters:
  • dirty_channel (numpy.ndarray) -- Dirty channel data

  • psf (numpy.ndarray) -- PSF for the given channel

  • nx (int) -- Number of x pixels

  • ny (int) -- Number of y pixels

  • cell_size (float) -- Cell size of each pixel in the image

  • algorithm (str) -- Cleaning algorithm: 'multiscale'|'iuwt'|'more_sane'|'generic_clean'

  • niter (int) -- Maximum number of iterations

  • threshold (float) -- Clean threshold

  • gain (float) -- loop gain (float) 0.7

  • scales (List[int]) -- Scales (in pixels) for multiscale ([0, 3, 10, 30])

Returns:

Tuple[numpy.ndarray, numpy.ndarray]

ska_sdp_spectral_line_imaging.data_procs.deconvolution.radler.radler_deconvolve_cube(dirty, psf, **kwargs)[source]

Note: This documentation copied from ska_sdp_func_python.image.deconvolution.radler_deconvolve_list

Clean using the Radler module, using various algorithms.

The algorithms available are (see: https://radler.readthedocs.io/en/latest/tree/cpp/algorithms.html)

multiscale iuwt more_sane generic_clean

For example:

comp, res = radler_deconvolve_cube(dirty_image, psf_image, nx=256, ny=256, cell_size=1.0, niter=1000, gain=0.7, algorithm='multiscale', scales=[0, 3, 10, 30], threshold=0.01)

Parameters:
  • dirty (Image) -- Cube dirty image

  • psf (Image) -- Point spread function image cube

  • **kwargs (keyword arguments) -- Additional keyword arguments

Return type:

Tuple[Image, Image]

Returns:

Tuple[Image, Image]

Component image cube, Residual Image cube