Version 1.9 Build 1556

News	FAQ
Search	Home

Next: MeasurementSet Classes Up: The Dwingeloo `UVCI' Design AIPS++ Implementation Note 181 Previous: Introduction

The Design in Brief

All the telescope data is stored in a MeasurementSet, which is a class derived from Table¹. The MeasurementSet classes differ from the Table class only in that:

1.: Methods are available to maintain a default correction (calibration) ``attached to'' the MeasurementSet.
2.: Some keywords and columns, if they exist in the MeasurementSet, must have predefined meanings. Moreover a MeasurementSet for a synthesis radiotelescopes is required to have some columns and keywords. Likewise, a Single Dish is required to have some columns and keywords, some of which are in common with the synthesis radiotelescopes.

The data in the MeasurementSet might be uncorrected (``raw'') data, or it might be (partially) corrected.

Such corrections (or corruptions) are applied to the MeasurementSet through the apply member of a Corrector object. Such corrections will typically not result in a physical copy of the data, rather they take place ``on demand.''

The machinery to perform this calibration on demand fundamentally uses the ``virtual column'' machinery available through the Table classes. When correction can be applied solely through the members available through a Visibility object, and the private state of the Corrector, even those details are largely hidden.

A Corrector may save and restore itself into any MeasurementSet keyword. The MeasurementSet has a keyword which is used to store a default calibration.

Image-plane data are handled either by the Image class, which contains regularly sampled data in a Lattice with associated coordinates, or a SkyModel, which represents a continuously sampled (in both the Image and Fourier planes) model of the sky brightness as a function of position, frequency, time and polarization. For now, a Beam/PSF is merely stored as an Image.

The MeasurementModel class models the processes by which a telescope samples the sky. Its invert method produces an Image and a Beam from a MeasurementSet, and its predict method produces a MeasurementSet from a SkyModel. The latter requires a MeasurementSet as input, to specify for which uvw-points the visibilities must be predicted.

Next: MeasurementSet Classes Up: The Dwingeloo `UVCI' Design AIPS++ Implementation Note 181 Previous: Introduction Contents