Hi Juan,
- If you are registering three or more frames that will be integrated later, then those dark pixels will be rejected during integration (provided you use a pixel rejection algorithm). In this case, just don't care. Keep an eye on the integrated image, however.
Just a note:
I have a situation where integration does not always reject these Lanczos-3 introduced dark pixels (with default clamping 0.3). The cases are integrations of 6 subs using percentile clipping (PC) due to the small number of subs. I see situations where a sufficient number of subs have dark pixels in the same location along the edges of stars that they are not rejected, at least with nominal PC rejection parameters. Subs were dithered and binned 1x1 with a FWHM of about 1.5 pixels.
Registration with bicubic spline clamping 0.1 does seem to avoid the problem.
Update: As a test I compared Lanczos-3 clamping 0.3 to Lanczos-3 clamping 0.1. The dark pixels remain present with 0.1, but they are less dark (brighter) than those of 0.3. On the other hand, both bicubic spline clamping 0.1 and 0.3 do a better job at avoiding them completely.
Thanks,
Mike
