mult-band processing

Here are some features that could be useful for multi-band processing:

• running sextractor in dual image mode so that that the lensing band can be used for object detection on the other bands
• some way of dealing with cases where coadds are not pixel matched in different bands
• integrating color information into deblender module (where needed)
• we need some connection to an external redshift code so that we can calculate the shear response of redshift dependent object selection (this involves some decisions on the scope of this pipeline. a minimal option would be to output a flux catalog with five columns (sheared + un-sheared flux) for each object, and leave it to the user to compute redshifts)
• the option to stack multiple bands for running shape measurement ( similar to DES)
• photometry options: do we want to fit the bands separately or do forced photometry etc
• we need some way to deal with zero point calibration, a minimal option could be to include a config file with externally computed zero points. (I personally found that these did not significantly change when using sheared and un-sheared fluxes, so one set is probably fine)

The issue names no files, tests, or entry points. Start by mapping the existing Python processing flow and turning the listed capabilities into separately scoped requirements; done would require an agreed scope and acceptance checks for the selected multi-band processing behavior.