tsdimaging

tsdimaging(infiles, outfile='', overwrite=False, field='', spw='', antenna='', scan='', intent='OBSERVE_TARGET#ON_SOURCE', timerange='', outframe='', mode='channel', nchan=-1, start='0', width='1', veltype='radio', specmode='cube', pointingcolumn='direction', convertfirst='never', projection='SIN', imsize='', cell='', phasecenter='', stokes='I', gridfunction='BOX', convsupport=-1, truncate='-1', gwidth='-1', jwidth='-1', clipminmax=False, minweight=0.1, brightnessunit='', restfreq='')[source]

SD task: imaging for total power and spectral data

[Description] [Examples] [Development] [Details]

Parameters
  • infiles (pathVec) - a list of names of input SD Measurementsets (only MS is allowed for this task)

  • outfile (string=’’) - prefix of output images (.image, .weight)

  • overwrite (bool=False) - overwrite the output file if already exists [True, False]

  • field ({string, stringVec}=’’) - select data by field IDs and names, e.g. '3C2*' (''=all)

  • spw ({string, stringVec}=’’) - select data by IF IDs (spectral windows), e.g. '3,5,7' (''=all)

  • antenna ({string, stringVec}=’’) - select data by antenna names or IDs, e.g, 'PM03' ('' = all antennas)

  • scan ({string, stringVec}=’’) - select data by scan numbers, e.g. '21~23' (''=all)

  • intent ({string, stringVec}=’OBSERVE_TARGET#ON_SOURCE’) - select data by observational intent, e.g. 'ON_SOURCE' (''=all)

  • timerange (string=’’) - select data by range of time, e.g. 20:15:00~20:16:00 (''=all)

  • mode (string=’channel’) - spectral gridding type ['channel', 'frequency', 'velocity']

    mode = channel
    • nchan (int=-1) - number of channels (planes) in output image (-1=all)

    • start ({string, int}=’0’) - start of output spectral dimension, e.g. '0', '110GHz', '-20km/s'

    • width ({string, int}=’1’) - width of output spectral channels

    mode = frequency
    • nchan (int=-1) - number of channels (planes) in output image (-1=all)

    • start ({string, int}=’0’) - start of output spectral dimension, e.g. '0', '110GHz', '-20km/s'

    • width ({string, int}=’1’) - width of output spectral channels

    mode = velocity
    • nchan (int=-1) - number of channels (planes) in output image (-1=all)

    • start ({string, int}=’0’) - start of output spectral dimension, e.g. '0', '110GHz', '-20km/s'

    • width ({string, int}=’1’) - width of output spectral channels

    • veltype (string=’radio’) - velocity definition ['radio', 'optical', 'true' or 'relativistic']

  • specmode (string=’cube’) - Spectral definition mode (cube, cubedata, cubesource)

    specmode = cube
    • outframe (string=’’) - velocity frame of output image ['lsrk', 'lsrd', 'bary', 'geo', 'topo', 'galacto', 'lgroup', 'cmb'] (''=current frame or LSRK for multiple-MS inputs)

  • pointingcolumn (string=’direction’) - pointing data column to use ['direction', 'target', 'pointing_offset', 'source_offset' or 'encoder']

  • convertfirst (string=’never’) - pointing column: direction conversion-interpolation processing scheme to use [“never”, “auto”, “always”]

  • projection (string=’SIN’) - map projection type

  • imsize ({intVec, doubleVec}=’’) - x and y image size in pixels, e.g., [64,64]. Single value: same for both spatial axes ([] = number of pixels to cover whole pointings in MSes)

  • cell ({string, stringVec, doubleVec}=’’) - x and y cell size, (e.g., ['8arcsec','8arcsec']. default unit arcmin. (’’ = 1/3 of FWHM of primary beam)

  • phasecenter (variant=’’) - image center direction: position or field index or ephemeris source info, e.g., 'J2000 17:30:15.0 -25.30.00.0', 'MARS'. ('' = the center of pointing directions in MSes)

  • stokes (string=’I’) - stokes parameters or polarization types to image, e.g. 'I', 'XX'

  • gridfunction (string=’BOX’) - gridding function for imaging ['BOX', 'SF', 'PB', 'GAUSS' or 'GJINC'] (see description in help)

    gridfunction = SF
    • convsupport (int=-1) - convolution support for gridding

    gridfunction = sf
    • convsupport (int=-1) - convolution support for gridding

    gridfunction = GAUSS
    • truncate ({string, int, double}=’-1’) - truncation radius for gridding

    • gwidth ({string, int, double}=’-1’) - HWHM for gaussian

    gridfunction = gauss
    • truncate ({string, int, double}=’-1’) - truncation radius for gridding

    • gwidth ({string, int, double}=’-1’) - HWHM for gaussian

    gridfunction = GJINC
    • truncate ({string, int, double}=’-1’) - truncation radius for gridding

    • gwidth ({string, int, double}=’-1’) - HWHM for gaussian

    • jwidth ({string, int, double}=’-1’) - c-parameter for jinc function

    gridfunction = gjinc
    • truncate ({string, int, double}=’-1’) - truncation radius for gridding

    • gwidth ({string, int, double}=’-1’) - HWHM for gaussian

    • jwidth ({string, int, double}=’-1’) - c-parameter for jinc function

  • clipminmax (bool=False) - Clip minimum and maximum value from each pixel

  • minweight (double=0.1) - Minimum weight ratio to use

  • brightnessunit (string=’’) - Overwrite the brightness unit in image ('' = respect the unit in MS) ['K' or 'Jy/beam']

  • restfreq ({string, double}=’’) - rest frequency to assign to image, e.g., '114.5GHz'. When the default value '' is used, the task currently internally uses REST_FREQUENCY of SOURCE table or mean frequency of given spectral window as a default value.

Description

The tsdimaging task grids/images total power and spectral data according to a specified gridding kernel. The input data should be calibrated and bandpass corrected (where necessary), and the output is a CASA image format dataset, either 2-D, 3-D, or 4-D depending on the input parameters. The details of conformance checks that are performed on the list of MSes are summarized in the CASA Docs page on Combining Datasets.

The output image contains up to four axes: two spatial axes, frequency, and polarization. By default, the spatial coordinates are determined such that the imaged area is covered with a cell size equal to 1/3 of the FWHM of the primary beam of antennas in the first MS. It is also possible to define the spatial axes of the image by specifying the image center direction (phasecenter), the number of image pixels (imsize), and the pixel size (cell).

The frequency coordinate of the image is defined by three parameters: the number of channels (nchan), the channel number/frequency/velocity of the first channel (start), and the channel width (width). The start and width parameters can be in units of ‘channel’ (use channel number), ‘frequency’ (e.g., ‘GHz’), or ‘velocity’ (e.g., ‘km/s’). By default, nchan, start, and width are set so that all selected spectral windows are covered with a channel width equal to the separation of the first two channels selected.

Finally, the polarization axis of the image is determined by the stokes parameter. For example, stokes=’XXYY’ produces an image cube with each plane containing the image of one of the polarizations, while stokes=’I’ produces a “total intensity”, or Stokes I image. There is also another option for Stokes I imaging, called ‘pseudoI’; the difference between ‘I’ and ‘pseudoI’ is how the task handles flag information. The stokes=’I’ imaging is stricter in the sense that it only takes into account visibilities for which all correlations are valid. In other words, it excludes all correlations for any data with any correlation flagged, even though the remaining correlations are valid. On the other hand, the ‘pseudoI’ option allows Stokes I images to include data for which either of the parallel hand data are unflagged.

Note

NOTE: Users should set stokes=’pseudoI’ if you want to get the equivalent result to the one obtained by setting stokes=’I’ for sdimaging. Setting stokes=’I’ in sdimaging is implemented the same way as stokes=’pseudoI’ in tsdimaging.

The parameter gridfunction sets the gridding function (convolution kernel) for imaging. Currently, the task supports ‘BOX’ (boxcar), ‘SF’ (Prolate Spheroidal Wave Function), ‘GAUSS’ (Gaussian), ‘GJINC’ (Gaussian*Jinc), where Jinc(x) = \(J_1(π*x/c)/(π*x/c)\) with a first order Bessel function J_1, and ‘PB’ (Primary Beam).

There are four subparameters for gridfunction: convsupport, truncate, gwidth, and jwidth. The convsupport parameter is an integer specifying the cutoff radius for ‘SF’ in units of pixels. By default (convsupport =-1), the cutoff radius is 3 pixels. The truncate parameter is a cutoff radius for ‘GAUSS’ or ‘GJINC’. It accepts integer, float, and string values, where the string would be a number plus unit. Allowed units include ‘deg’, ‘arcmin’, ‘arcsec’, and ‘pixel’. The default is ‘pixel’. The default value for truncate, which is used when a negative radius is set, is 3*HWHM for ‘GAUSS’ and the radius at the first null for ‘GJINC’. The gwidth is the HWHM of the Gaussian for ‘GAUSS’ and ‘GJINC’. The default value is \(sqrt(log(2))\) pixels for ‘GAUSS’ and \(2.52*sqrt(log(2))\) pixels for ‘GJINC’. The jwidth specifies the width of the jinc function (parameter ‘c’ in the definition above). The default is 1.55 pixels. Both gwidth and jwidth allow integer, float, or string values, where the string would be a number plus unit. The default values for gwidth and jwidth are taken from Mangum, et al. 2007 [1]. The formula for ‘GAUSS’ and ‘GJINC’ are taken from Table 1 in the paper, and are written as follows using gwidth and jwidth:

GAUSS: \(\exp[-\log(2)*(|r|/gwidth)^2]\),

GJINC: \(J_1(π*|r|/jwidth)/(π*|r|/jwidth)* \exp[-\log(2)*(|r|/gwidth)^2]\),

where \(r\) is the radial distance from the origin.

The outfile should be unique. tsdimaging will stop with an Exception error (e.g., Exception: Unable to open lattice) if outfile is the same as the infiles name.

The phasecenter parameter sets the center direction of the image. If the phasecenter is the name known major solar system object (‘MERCURY’, ‘VENUS’, ‘MARS’, ‘JUPITER’, ‘SATURN’, ‘URANUS’, ‘NEPTUNE’, ‘PLUTO’, ‘SUN’, ‘MOON’) or is an ephemerides table then that source is tracked and the background sources get smeared. A parameter ephemsrcname has been deprecated accordingly. There is a special case, when phasecenter=’TRACKFIELD’, which will use the ephemerides or polynomial phasecenter in the FIELD table of the MeasurementSets as the source center to track.

The clipminmax function can clip minimum and maximum values from each pixel. This function makes the computed output slightly more robust to noise and spurious data. Note that the benefit of clipping is lost when the number of integrations contributing to each gridded pixel is small, or where the incidence of spurious data points is approximately equal to or greater than the number of beams (in area) encompassed by the expected image.

The minweight parameter defines a threshold of weight values to mask. The pixels in outfile whose weight is smaller than minweight *median (weight) are masked out. The task also creates a weight image with the name outfile.weight.

The projection parameter allows to specify what kind of map projection is applied. The default is SIN (slant orthographic) projection. The task also supports CAR (plate carrée), TAN (gnomonic), and SFL (Sanson-Flamsteed).

Image Definition (specmode)

The image coordinate system(s) and shape(s) can be set up to form single images. The different modes for imaging include:

  • ‘cube’: Spectral line imaging with one or more channels. The fixed spectral frame, LSRK, will be used for automatic internal software Doppler tracking so that a spectral line observed over an extended time range will line up appropriately. You can change the output spectral frame via outframe parameter.

  • ‘cubedata’: Spectral line imaging with one or more channels There is no internal software Doppler tracking so a spectral line observed over an extended time range may be smeared out in frequency.

  • ‘cubesource’: Spectral line imaging while tracking moving source (near field or solar system ephemeris objects). The velocity of the source is accounted and the frequency reported is in the source frame.

Technical Note: sdimaging and tsdimaging

The tsdimaging task is supposed to replace sdimaging. The initial version of this task was intended to be fully compatible with sdimaging. Technically speaking, those tasks share underlying framework with interferometry imaging tasks: sdimaging shares with clean, while tsdimaging is based on the framework for tclean. As clean (and the underlying framework) will be deprecated and replaced with tclean in the future, sdimaging will also be made obsolete in favor of migrating to tsdimaging. This transition will have several benefits from the user’s point of view in future. In terms of functionality, new features implemented in tclean will also apply to tsdimaging if the features are useful for single dish imaging. Another possible benefit is a performance. Since the framework for tclean is designed to support parallel processing, it can also be used to speed up tsdimaging. This should be effective for large datasets, but these examples represent future work. Currently, effort is underway to make tsdimaging compatible with sdimaging and convert it to a “regular” (non-experimental) task.

Bibliography

Examples

To generate a spectral line cube with 500 channels selected from channel 200 to 700:

spw='0'
pol='XX'
src='Moon'

tsdimaging(infiles='mydata.ms',
           spw=spw,
           nchan=500,
           start='200',
           width='1',
           cell=['30.0arcsec','30.0arcsec'],
           outfile='mydata.ms.im',
           imsize=[80,80],
           gridfunction='GAUSS',
           gwidth='4arcsec',
           stokes=pol,
           ephemsrcname=src)

The start parameter can be specified in different units:

start=100  # mode='channel'
start='22.3GHz'  # mode='frequency'
start='5.0km/s'  # mode='velocity'

The parameter phasecenter sets the center direction of the image:

phasecenter=6
phasecenter='J2000 19h30m00 -40d00m00'
phasecenter='J2000 292.5deg -40.0deg'
phasecenter='J2000 5.105rad -0.698rad'
phasecenter='ICRS 13:05:27.2780 -049.28.04.458'
phasecenter='myComet_ephem.tab'
phasecenter='MOON'
phasecenter='TRACKFIELD'
Development

No additional development details

Parameter Details

Detailed descriptions of each function parameter

infiles (pathVec) - a list of names of input SD Measurementsets (only MS is allowed for this task)
outfile (string='') - prefix of output images (.image, .weight)
overwrite (bool=False) - overwrite the output file if already exists [True, False]
field ({string, stringVec}='') - select data by field IDs and names, e.g. '3C2*' (''=all)
spw ({string, stringVec}='') - select data by IF IDs (spectral windows), e.g. '3,5,7' (''=all)
antenna ({string, stringVec}='') - select data by antenna names or IDs, e.g, 'PM03' ('' = all antennas)
scan ({string, stringVec}='') - select data by scan numbers, e.g. '21~23' (''=all)
intent ({string, stringVec}='OBSERVE_TARGET#ON_SOURCE') - select data by observational intent, e.g. 'ON_SOURCE' (''=all)
timerange (string='') - select data by range of time, e.g. 20:15:00~20:16:00 (''=all)
outframe (string='') - velocity frame of output image (‘’=current frame or LSRK for multiple-MS inputs)
mode (string='channel') - spectral gridding type ['channel', 'frequency', 'velocity']
nchan (int=-1) - number of channels (planes) in output image (-1=all)
start ({string, int}='0') - start of output spectral dimension, e.g. '0', '110GHz', '-20km/s'
width ({string, int}='1') - width of output spectral channels
veltype (string='radio') - velocity definition
specmode (string='cube') - Spectral definition mode (cube, cubedata, cubesource)
pointingcolumn (string='direction') - pointing data column to use
convertfirst (string='never') - Specify whether the direction of the specified pointing column must be converted to image”s reference frame prior to being interpolated at data-taking time, and when. “never”: interpolate against the pointing column, then convert. “always”: interpolate against the beforehand converted pointing column. “auto”: if there are less pointings than selected data rows convert first, else interpolate first
projection (string='SIN') - map projection type
imsize ({intVec, doubleVec}='') - x and y image size in pixels, e.g., [64,64]. Single value: same for both spatial axes ([] = number of pixels to cover whole pointings in MSes)
cell ({string, stringVec, doubleVec}='') - x and y cell size, (e.g., ['8arcsec','8arcsec']. default unit arcmin. (’’ = 1/3 of FWHM of primary beam)
phasecenter (variant='') - image center direction: position or field index or ephemeris source info, e.g., 'J2000 17:30:15.0 -25.30.00.0', 'MARS'. ('' = the center of pointing directions in MSes)
stokes (string='I') - stokes parameters or polarization types to image, e.g. 'I', 'XX'
gridfunction (string='BOX') - gridding function for imaging (see description in help)
convsupport (int=-1) - convolution support for gridding
truncate ({string, int, double}='-1') - truncation radius for gridding
gwidth ({string, int, double}='-1') - HWHM for gaussian
jwidth ({string, int, double}='-1') - c-parameter for jinc function
clipminmax (bool=False) - Clip minimum and maximum value from each pixel. Note the benefit of clipping is lost when the number of integrations contributing to each gridded pixel is small, or where the incidence of spurious datapoints is approximately or greater than the number of beams (in area) encompassed by expected image.
minweight (double=0.1) - Minimum weight ratio to the median of weight used in weight correction and weight beased masking
brightnessunit (string='') - Overwrite the brightness unit in image ('' = respect the unit in MS) ['K' or 'Jy/beam']
restfreq ({string, double}='') - rest frequency to assign to image, e.g., '114.5GHz'. When the default value '' is used, the task currently internally uses REST_FREQUENCY of SOURCE table or mean frequency of given spectral window as a default value.