FXE asked us to extend the ROI projection we already integrated with a threshold, zeroing out pixels below that threshold before reducing the ROI to a 1D curve. This gets rid of a lot of remaining noise (see the screenshot below).
How Has This Been Tested?
Ran the notebook interactively on p3279 r275, with ROI definition [3, 405, 450, 0, 1024, -2] (supplied by Yohei), with no thresholding and with the threshold set to 1.
Relevant Documents (optional)
Average of the projected ROI spectrum over the whole run, with / without thresholding:
Types of changes
New feature (non-breaking change which adds functionality)
