Image Combination

When you are satisfied with the deconvolution, restoration and self-calibration of all the individual images, task linmos can be used to combine them in a linear mosaic. Usually you will just combine the restored images (if you are going to do quantitative analysis on the composite image, it may be best to do a deep CLEAN and use the same restoring beam for all pointings). Although linmos can interpolate input images to align them, its algorithm, particularly for geometric correction, is very poor, and so this is strongly discouraged. You should use invert to make all the input images on the same grid, by setting a common tangent point (offset keyword).

Task linmos uses the same weighted sum of the input pointings as the `joint approach' software (see Section 21.6). Normally the expected rms noise in the image is determined from the images themselves (image item rms). However if this item is missing, or if you wish to override it to get a different weighting, you may enter the expected rms noise via keyword rms. Also note that linmos, by default, fully corrects for the primary beam attenuation even when this excessively amplifies the noise. The taper option can be used to reduce the correction at the edge of the field, and thus avoid excessive noise amplification.

Typical inputs to linmos are:

LINMOS
in=lmc_*.cln	Use wildcards to select all images.
out=lmc.mos	The output linearly mosaiced image.
rms	Generally left unset.
options	Leave blank to fully correct primary beam,
options=taper	or set to taper at the edge of the mosaic.