Mosaic imaging

NRAO Home > CASA > CASA Cookbook and User Reference Manual

5.3.15 Mosaic imaging

The clean task contains the capability to image multiple pointing centers together into a single “mosaic” image. This ability is controlled by setting imagermode=’mosaic’ (§ 5.3.4).

The key parameter that controls how clean produces the mosaic is the ftmachine sub-parameter (§ 5.3.4.3). For ftmachine=’ft’, clean will perform a weighted combination of the images produced by transforming each mosaic pointing separately. This can be slow, as the individual sub-images must be recombined in the image plane. NOTE: this option is preferred for data taken with sub-optimal mosaic sampling (e.g. fields too far apart, on a sparse irregular pattern, etc.).

The primary beams used in CASA are described in § 5.2.13.

If ftmachine=’mosaic’, then the data are gridded onto a single uv-plane which is then transformed to produce the single output image. This is accomplished by using a gridding kernel that approximates the transform of the primary beam pattern. Note that for this mode the <imagename>.flux image includes this convolution kernel in its effective weighted response pattern (needed to “primary-beam correct” the output image). For this mode only, an additional image <imagename>.flux.pbcoverage is produced that is the primary-beam coverage only used to compute the minpb cutoff (§ 5.3.7).

ALERT: In order to avoid aliasing artifacts for ftmachine=’mosaic’ in the mosaic image, due to the discrete sampling of the mosaic pattern on the sky, you should make an image in which the desired unmasked part of the image (above minpb) lies within the inner quarter. In other words, make an image twice as big as necessary to encompass the mosaic.

It is also important to choose an appropriate phasecenter for your output mosaic image (§ section:im.pars.phasecenter). The phase center should not be at the edge of an image with pointings around it. In that case, FFT aliasing may creep into the image.

An example of a simple mosaic clean call is shown below:

clean(vis=’n4826_tboth.ms’,
       imagename=’tmosaic’,
       mode=’channel’,
       nchan=30,start=46,           # Make the output cube 30 chan
       width=4,                     # start with 46 of spw 0, avg by 4 chans
       spw=’0~2’,
       field=’0~6’,
       cell=[1.,1.],
       imsize=[256,256],
       stokes=’I’,
       psfmode=’clark’,
       niter=500,
       imagermode=’mosaic’,
       scaletype=’SAULT’,
       cyclefactor=0.1)

Another example of mosaic imaging, this time using 3mm BIMA data, is given in Appendix F.3.

[next] [prev] [prev-tail] [front] [up]

More information about CASA may be found at the CASA web page

This code is available under the terms of the GNU General Public Lincense