Version 1.9 Build 1556

News	FAQ
Search	Home

Next: Simulation Up: The Dwingeloo `UVCI' Design AIPS++ Implementation Note 181 Previous: MeasurementSet Correction (Calibration) and Corruption

MeasurementModel

A MeasurementModel is an abstract base class which encapsulates (an approximation to) the process by which a telescope samples the sky. The fundamental members of the MeasurementModel class are:

invert: The invert() method creates an Image and Beam (presently also an Image) from a MeasurementSet. The present strategy is that the MeasurementSet must have previously had any desired selections applied to it (e.g., for field, or frequencies). An alternative strategy would be to put these filters into the MeasurementModel.
predict: The predict() method returns a MeasurementSet, given a SkyModel and a template MeasurementSet. Predict uses the template ``coordinates'' (e.g., UVW) to determine where to sample the model.

A class derived from MeasurementModel that implements the usual aperture synthesis image formation has been implemented. It handles spectral line and continuum data and is also capable of performing multi-frequency synthesis. It has options for 3D imaging, tangent-point shifts and emitting double-sized beams (which are often useful for image-plane deconvolutions).

Without a more generalized image structure, we are unsure whether or not the MeasurementModel can be truly polymorphic. For example, a mosaicing MeasurementModel would certainly need to produce a class more complicated than an Image.

Next: Simulation Up: The Dwingeloo `UVCI' Design AIPS++ Implementation Note 181 Previous: MeasurementSet Correction (Calibration) and Corruption Contents