Open hjkristenson opened 4 years ago
asjohnston-asf
commented
3 years ago Additionally, we should provide a default threshold value (-24?) for rtc2color.py and rtc2color() so novice users (like myself) have at least one example value rather than making a wild guess.
https://github.com/ASFHyP3/hyp3-lib/blob/develop/hyp3lib/rtc2color.py#L125 https://github.com/ASFHyP3/hyp3-lib/blob/develop/hyp3lib/rtc2color.py#L210
tomalogan
commented
3 years ago Talked with Rudi.
The numbers for HyP3 were empirically determined. That is, Jeremy spent quite some time processing images and examining the colors, adjusting the threshold value. The three numbers (4) of -19, -22.5 (-23.5), and -24 were each picked because they made some images easier to interpret. The -23.5 is for EW, which Rudi said definitely had a different number than IW.
The reason for the different numbers is the variability of water's backscatter.
Since -22.5 (-23.5) are in the middle, I would propose that we use -22.5 as the default. Maybe Kirk can weigh in on why we chose -24 in the current code. I don't know.
Finally, Rudi indicated that the teal option is good for glaciers specifically. Thus, it is valuable and should be retained.
hjkristenson
commented
3 years ago I ended up sticking with the -24 dB default value for the ArcGIS Toolbox, and that works well for IW scenes in general. I find that the larger values generate more blue on the landscape than would be expected. It's also consistent with the value we've been using for RGB products and browses generated in HyP3, and in this case I think consistency is a reasonable path to follow. It's all what you get used to looking at.
talogan
commented
3 years ago OK. -24 sounds good to me! Let's go with that for default.
On Mon, Dec 14, 2020 at 6:32 PM Heidi Kristenson notifications@github.com wrote:
I ended up sticking with the -24 dB default value for the ArcGIS Toolbox, and that works well for IW scenes in general. I find that the larger values generate more blue on the landscape than would be expected. It's also consistent with the value we've been using for RGB products and browses generated in HyP3, and in this case I think consistency is a reasonable path to follow. It's all what you get used to looking at.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/ASFHyP3/hyp3-lib/issues/210#issuecomment-745029683, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAP2NOKNGCVFQVFX36IWVTLSU3KENANCNFSM4RST7REQ .
In the documentation we write that a suitable threshold range is -22 to -25 (I think that came from Jeremy's doc, but I don't remember for sure), but in HyP3 the options we provide are -19, -22.5 (-23.5 for IW), -24.
We should figure out the source of the differences. Is the -19 option for a specific dataset or acquisition mode (i.e. one of the stripmap beam modes)? And what is being differentiated with the second option? Is that EW vs IW (and do we really apply -22.5 for one and -23.5 for the other)?