DM-51111: Fix to ITL saturation sag masking and add log#417
DM-51111: Fix to ITL saturation sag masking and add log#417
Conversation
aferte
changed the title
aferte
changed the title
erykoff
left a comment
erykoff
left a comment
There was a problem hiding this comment.
I will ask again, as I did on the ticket. Is it possible to check the orientation of the saturated footprint? The sag masking should only trigger on vertical streaks, and right now it just looks at footprint size and not the aspect ratio of the footprint. Have you looked into this at all?
aferte
force-pushed
the
tickets/DM-51111
branch
from
144923e to
d6cd5b8
Compare
Add check on shape of saturated box. Add requirement on footprint shape. Update footprint ratio threshold and add log info.
aferte
force-pushed
the
tickets/DM-51111
branch
from
d6cd5b8 to
c95eb08
Compare
|
@erykoff thanks - here is an update adding a check on the shape of the saturated footprint. Jenkins is running now with that update |
| maskedImage.mask.array[:, columnsToMask] |= saturatedBit | ||
| if ratioSatFP > 0.5: | ||
| # We apply the masking if the saturation footprint has a width of | ||
| # at most twice the height, to avoid trigerring in the case of |
There was a problem hiding this comment.
Is this correct? "width at most twice the height"?
There was a problem hiding this comment.
yes, if the width is more than twice the height the saturation sag masking won't be applied
There was a problem hiding this comment.
My argument is that a saturation streak (that might sag) will always be taller than wide. Is there evidence this is not true?
There was a problem hiding this comment.
It can happen that the saturation streak is within a footprint with an edge bleed - if the source is close to the edge, the footprint (e.g. edge bleed + source + saturation sag) can be wider than tall because of the edge bleed.
2 participants
No description provided.