Optimised post-exposure image sharpening code for L3-CCD detectors

As light from celestial bodies traverses Earth's atmosphere, the wavefronts are distorted by atmospheric turbulence, thereby lowering the angular resolution of ground-based imaging. Rapid time-series imaging enables Post-Exposure Image Sharpening (PEIS) techniques, which employ shift-and-add frame registration to remove the tip-tilt component of the wavefront error - as well as telescope wobble, thus benefiting all observations. Further resolution gains are possible by selecting only frames with the best instantaneous seeing - a technique sometimes calling "Lucky Imaging". We implemented these techniques in the 1990s, with the TRIFFID imaging photon-counting camera, and its associated data reduction software. The software was originally written for time-tagged photon-list data formats, recorded by detectors such as the MAMA. This paper describes our deep re-structuring of the software to handle the 2-d FITS images produced by Low Light Level CCD (L3-CCD) cameras, which have sufficient time-series resolution (>30Hz) for PEIS. As before, our code can perform straight frame co-addition, use composite reference stars, perform PEIS under several different algorithms to determine the tip/tilt shifts, store "quality" and shift information for each frame, perform frame selection, and generate exposure-maps for photometric correction. In addition, new code modules apply all "static" calibrations (bias subtraction, dark subtraction and flat-fielding) to the frames immediately prior to the other algorithms. A unique feature of our PEIS/Lucky Imaging code is the use of bidirectional wiener-filtering. Coupled with the far higher sensitivity of the L3-CCD over the previous TRIFFID detectors, much fainter reference stars and much narrower time windows can be used.