• Ultraviolet Photography
  •  

Snow in the SWIR redux

Infrared SWIR Multispectral
14 replies to this topic

#1 Andy Perrin

    Member

  • Members+G
  • 4,084 posts
  • Location: United States

Posted 21 December 2020 - 21:59

I finally redid my pano today, thanks to the overcast weather. (Link to previous attempt) When the sun is out, my mere 10 bits of dynamic range are not enough to capture everything in the photo. It becomes like trying to photograph the moon AND the stars simultaneously -- pick one!

Even so, there was enough variation in the photo that I had some trouble stitching the bottom edge with all that black snow. The snow seemed much darker today, which is probably a sign of the thaw, since it was well above 0 deg C.

Details are the same as the last attempt except that I stitched in Hugin.

Attached Image: pano3_blended_fused UVP.jpg

Visible (today)
Attached Image: pano3 visible UVP.jpg

Previous pano for comparison purposes. We have had additional snow + melting in the interim.
Attached Image: Snow SWIR UVP first try.jpg

Edited by Andy Perrin, 22 December 2020 - 02:55.


#2 Stefano

    Member

  • Members(+)
  • 1,796 posts
  • Location: Italy

Posted 21 December 2020 - 22:06

Andy, very nice, much better than your first attempt. Seeing those black stacks of snow in the street edges/corners is so nice and odd. The world really is a different place in SWIR.

#3 Bernard Foot

    Bernard Foot

  • Members+G
  • 718 posts
  • Location: UK

Posted 21 December 2020 - 22:32

The black snow is staggering. Looks like someone has ben dumping coal on the street. The white house is looking cleaner, though.
Bernard Foot

#4 Andy Perrin

    Member

  • Members+G
  • 4,084 posts
  • Location: United States

Posted 21 December 2020 - 22:46

Bernard, I wish all my stains disappeared the way the ones on that white (yellow) house did!

#5 Andy Perrin

    Member

  • Members+G
  • 4,084 posts
  • Location: United States

Posted 21 December 2020 - 23:10

Here is the 980nm version.
Attached Image: snow pano 980nm UVP_res.jpg

Repeating the above 1500nm version here for comparison.
Attached Image: pano3_blended_fused UVP.jpg

And the ~10000nm version (taken during daytime, so not strictly thermal IR, just LWIR). This was taken the day after the first storm, so not quite identical to the above two.
Attached Image: pano4 crop_res UVP.jpg

Edited by Andy Perrin, 21 December 2020 - 23:16.


#6 dabateman

    Da Bateman

  • Members+G
  • 2,767 posts
  • Location: Maryland

Posted 22 December 2020 - 02:21

Andy the first post here seems to have the photos deleted. but the post above I see stuff so get an idea.

For that coal look need 1500nm, that 980 filter will not cut it.

#7 Andy Perrin

    Member

  • Members+G
  • 4,084 posts
  • Location: United States

Posted 22 December 2020 - 02:53

Weeeeird. I reuploaded the photos because they self-deleted earlier also. I wonder if our database problems aren't over yet. I will try to repair it now.
--

Yes, definitely you won't get black snow at 980nm. There is an additional water peak at 1205nm or so that I haven't tried yet. I wonder if you get gray snow?

Edited by Andy Perrin, 22 December 2020 - 02:57.


#8 dabateman

    Da Bateman

  • Members+G
  • 2,767 posts
  • Location: Maryland

Posted 22 December 2020 - 04:54

We had some yellow snow, but its mostly all melted now.

I see the photos now. Black snow is interesting. You seem to be at the edge of teasing out water vs ice vs snow with these bands. You could maybe try an odd test in a glass or plastic to get the layers.

#9 Andy Perrin

    Member

  • Members+G
  • 4,084 posts
  • Location: United States

Posted 22 December 2020 - 05:25

View Postdabateman, on 22 December 2020 - 04:54, said:

We had some yellow snow, but its mostly all melted now.

I see the photos now. Black snow is interesting. You seem to be at the edge of teasing out water vs ice vs snow with these bands. You could maybe try an odd test in a glass or plastic to get the layers.
Just don't eat the yellow kind.

With regard to your second comment, I would one day like to try a tri-color with 980nm, 1205nm, and 1450nm which are (in round numbers) the three water peaks. That might do the job. I don't have a 1205nm bandpass filter though.

#10 Stefano

    Member

  • Members(+)
  • 1,796 posts
  • Location: Italy

Posted 22 December 2020 - 05:44

View PostAndy Perrin, on 22 December 2020 - 05:25, said:

With regard to your second comment, I would one day like to try a tri-color with 980nm, 1205nm, and 1450nm which are (in round numbers) the three water peaks. That might do the job. I don't have a 1205nm bandpass filter though.
I'm so interested in this. For the 1205 nm peak a 1200 nm bandpass would work Ok, as that peak is also quite broad.

Would you use your TriWave for all shots, including the 980 nm one, to make images perfectly (or almost) stackable?

#11 Andy Perrin

    Member

  • Members+G
  • 4,084 posts
  • Location: United States

Posted 22 December 2020 - 05:56

Oh, it would have to be the TriWave for the reason you said. The resolution isn't the same on the other camera, and the sensor size is different, so there would need to be a lot of manipulation otherwise.

#12 dabateman

    Da Bateman

  • Members+G
  • 2,767 posts
  • Location: Maryland

Posted 22 December 2020 - 12:38

Andy,
Its just as bad to eat the grey/black kind. Especially if yellow is next to it.

Funny how my desire to image water on the other end (185nm) is just as hard as on your end at 1500nm. Although, we may both be there now. I still need a good light.

#13 Stefano

    Member

  • Members(+)
  • 1,796 posts
  • Location: Italy

Posted 22 December 2020 - 13:02

Sometimes SWIR reminds me of UVC. You need special lenses for UVC, and not all lenses are good in SWIR (most are, it seems). UVC and SWIR LEDs have similar efficiencies, power and prices. Skin is black in both bands, things start to absorb again in SWIR (and most things absorb UVC), the atmosphere is opaque to UVC and starts to be opaque in SWIR, and of course water absorbs well in both bands (especially below 200 nm, the absorption rises very quickly). Going far in both directions is difficult.

David, even a 10 nm decrement in wavelength would probably do a lot. Water started to appear dark at 185 nm, but by 165/170 nm it should be pitch black. Look here: https://en.m.wikiped...iquid_water.png

Also, you may want to use distilled water or as pure of a water as you can get. Down there even tiny amounts of impurities probably absorb a lot.

#14 Andy Perrin

    Member

  • Members+G
  • 4,084 posts
  • Location: United States

Posted 22 December 2020 - 19:19

I have ordered filters at the following wavelengths:
1200nm
1250nm (it was a group of two in the eBay listing, so this is a “freebie”)
1450nm

All are 10 to 12nm bandpass.

All are originally from Thorlabs so should be correctly labeled. 1450 is the actual water absorption peak.

These should allow head-to-head comparisons after I normalize with some kind of white standard.

#15 Stefano

    Member

  • Members(+)
  • 1,796 posts
  • Location: Italy

Posted 22 December 2020 - 20:47

Spectralon works well in SWIR too, so if you have some you can use that as your white standard.

The 1450 nm filter will be the most interesting. That’s the main wavelength that makes SWIR what it is. Black water is one of the main reasons why someone would want to reach that band.

I think it could become your standard SWIR filter if you are looking for water-related patterns or stains. Ice peaks higher, but water will be as dark as you can get it with your camera. The next stronger peak is at about 1900-1950 nm, outside your camera’s range.

Who knows, one day you may have a MWIR camera or a camera that can see in the 2000-3000 nm range. After about 3000 nm (this is a very arbitrary and broad cut-off) things start glowing, so darkness will be difficult to achieve. But you may catch the very strong ~2950 nm peak.

Edited by Stefano, 22 December 2020 - 21:21.