Common pixel grid for multi-filter calimage3 products, for a given observation

[stscijgbot-jp]

Issue JP-2296 was created on JIRA by Anton Koekemoer:

If an observation has images in multiple filters, these are passed to the pipeline as separate associations, and the processing is done separately for each filter. This is appropriate, and correct, for certain steps in the calimage3 pipeline, eg outlier rejection which needs to process data from only a single filter at a time.

However, at the point where the final mosaics are constructed, there is currently no mechanism to ensure that exposures obtained in different filters (in a given observation, within a given program) will be on the same pixel grid, since the output mosaic WCS is constructed independently for each filter, based on the layout of its exposures.

This therefore limits the usability of the calimage3 mosaics -- in general, when an observing program obtains images in multiple filters on a given location in the sky, the expectation is that science can be carried out on these different filters in mosaics that are aligned to the same pixel grid.

This ticket is therefore aimed at examining ways in which the output WCS, at least for multiple filters within a given observation, in a given program, can be set up in such a way that observations obtained in different filters can end up as mosaics that are on the same pixel grid, thereby significantly adding scientific value to the output products.