mult-band processing

[lbaumo]

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)

[martinkilbinger]

We have implemented the option to run SExtractor in dual-image mode. I have done this to get multi-band photometry. This was run on the tiles, but not on the single-exposure images. So not too useful so far for multi-epoch shape measurement.

[lbaumo]

Is the astrometric solution that was used to make the coadds available? The scamp solutions were sufficient for my data