My Starlight Express H16 has quite a few hot pixels that scaled darks does not adequately deal with (below bottom left). Many of these have adjacent warm or cold pixels so there are errant pixel clusters that Cosmetic Correction alone then fails to correct. When these frames are then aligned these little clusters tend to grow(below top right) , even if the clamping threshold is reduced considerably and whatever interpolation routine is used, though it can help. I tend to dither ~6 imager pixels (~10") and this does deal with them for the most part but on large data sets does perhaps require me to maintain more restrictive rejection thresholds than I might otherwise get away with and can sometime lead to odd artefacts even in high signal areas if the clusters interact.
I can clean the frame up significantly with a defect map (from a master dark) that deals with the main culprits and isolates the minors such that cosmetic correction then deals with them (below top left) with pretty permissive sigma values. Comparison shows a small but useful increase in stacked S/N as these pixels are no longer a cause for concern allowing free range on rejection criteria. So my current pre-processing workflow is
Image calibration (Bias, Scaled Darks, Flats)
Defect Map (Inverted Binarise of Master Dark threshold = 0.04)
Cosmetic Correction (Autocorrect Hot and Cold Sigma = 4.0)
Star Alignment (Default Settings)
Image Integration
The extra step of Defect Mapping using Image Container is hardly a major issue but I am wondering if I am missing a trick in dealing with the problem.
Chris
