This regards Averaged Sigma Clipping. In the in-app documentation, it says
"As we have implemented it, this algorithm works well for image sets of 10 or more images. The original implementation in IRAF works for smaller data sets because the proportionality between photons and pixel values, or the estimated sensor gain (the s variable in the algorithm above) is calculated for each row of pixels. In our implementation the sensor gain is estimated separately for each pixel stack, in order to avoid visible differences between adjacent rows in the output image."
But in the tooltip help for the Rejection Algorithm selector box in ImageIntegration, it says that it's a good algorithm for small sets of 3 to 10 images.
Based on reading both, I'm assuming that the in-app doc is more correct than the tooltip, but wanted to confirm that and also point out the inconsistency... Right now I'm dealing with 5-image sets and obviously want to pick the correct algorithm. Until I hear differently I'll be using Percentile Clipping for sets <8 images, and Winsorized for 8 or more.
Thanks,
Dave B.
