Data reduction strategies for lucky imaging

Lucky imaging is a proven technique for near diffraction limited imaging in the visible; however, data reduction and analysis techniques are relatively unexplored in the literature. In this paper we use both simulated and real data to test and calibrate improved guide star registration methods and noise reduction techniques. In doing so we have produced a set of "best practice" recommendations. We show a predicted relative increase in Strehl ratio of ~ 50% compared to previous methods when using faint guide stars of ~17th magnitude in I band, and demonstrate an increase of 33% in a real data test case. We also demonstrate excellent signal to noise in real data at flux rates less than 0.01 photons per pixel per frame above the background flux level.