Public Lab Research note


Canon SD4000IS as Infragram

by warren | August 08, 2014 18:50 | 2,005 views | 9 comments | #11023 | 2,005 views | 9 comments | #11023 08 Aug 18:50

Read more: unstable.publiclab.org/n/11023


What I want to do

We'd struggled a year ago to get a CMOS-sensor based Canon SD4000IS working as an Infragram modified camera, so when @nedhorning, @mathew and @cfastie posted about the red filter Infragram technique, I was eager to see if that could get this excellent camera working. You can read more about Mathew and Chris's research here: http://publiclab.org/notes/mathew/04-30-2014/rosco-replacements-for-wratten-25a-pt-2

My attempt and results

My SD4000IS already had its infrared block filter removed, so I just taped a piece of Rosco Fire filter over the front. It seems to have worked very well! Much better differentiation than with a blue filter:

http://infragram.org/i/53e519b10ab503231b000c17 (lead image)

http://infragram.org/i/53e519870ab503231b000c16

2014-08-08T18_47_44.804Z.jpg

http://infragram.org/i/53e519440ab503231b000c15

2014-08-08T18_48_11.946Z.jpg

Questions and next steps

It'd be more ideal to test these in a grassy field, but these are encouraging!


9 Comments

Your new Infragram SD4000 seems to have promise. Here is one with my new lut:
.
PottedNDVI.JPG

Reply to this comment...


Oh cool! do you want to post your LUT in a numeric format so we can post it as an Infragram.org option?

The format is:

colormap = segmented_colormap([[0, [0, 0, 255], [38, 195, 195]], [0.5, [0, 150, 0], [255, 255, 0]], [0.75, [255, 255, 0], [255, 50, 50]]]);

where the outermost [ ] contain a set of [x-position,[r,g,b]] items -- x-position measured from 0-1.

And a name for your LUT and a place to link to for more information about it. Maybe http://publiclab.org/wiki/infragram-luts ?

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


I will have to see if Fiji can output the values for a lut. Fiji has a GUI lut editor, so I have no idea what the values are. Maybe Ned can help. Here is my lut in the editor:
.

VGYRMlut.JPG
.
I was hoping to post a note on this lut. I just tried to attach the lut file to this comment but dragging it here seemed to make somebody mad.

Reply to this comment...


we only accept certain types of files for upload, for security reasons. I can add .lut, is it an image file or a text file?

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


I'm not sure what a Fiji lut file is. Might be binary. I got the values into a tab delimited text file.

NDVI_VGYRM-lut.txt

And I was able to circumvent the dastardly security measures to attach it here.

I guess it might be easy for certain people to make this usable at infragram.org. I wouldn't know.

But this is a very new lut and I am just starting to put it to use and evaluate it. So I will post a note about it and see if people have opinions.

Reply to this comment...


Whoa... no, a bit tough to use. If you can describe it as a sequence of colors, the format I mentioned will interpolate gradients between them. Just looking at yours, it seems like it could be:

0: white

0.167: black

0.33: white

0.5: black

0.55: blue

0.7: yellow

0.9: red

1.0: magenta

what do you think?

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


I get it now. Your translation is good. The blue is actually a light blue, I think [140 140 255] at 0.56. And I put pure green in there explicitly at 0.63 instead of deriving it from a gradient from light blue to yellow (at 0.75). Red is at 0.94.

Reply to this comment...


OK - try it out! I don't think it works in the WebGL version but in the basic version it should: http://infragram.org/sandbox/

Reply to this comment...


Good stuff.

Reply to this comment...


Login to comment.