why do plants reflect infrared?

[Jeffrey Warren]

I did a new diagram explaining why plants reflect so much infrared (and thus why NVI and NRG work). Feedback?

http://www.flickr.com/photos/jeffreywarren/5412520298/

I'm just having so much fun with these diagrams that I have to share.

[Cesar Harada]

This diagram is really really cute :)
Everything has a different response to light, plants, animals, us.
http://en.wikipedia.org/wiki/Phase_response_curve

The transformation if light into food is also a marvelous mechanism :
http://wiki.answers.com/Q/What_is_the_formula_for_photosynthesis

In fact not only land plants do photosynthesis, algae too :
http://en.wikipedia.org/wiki/Photosynthesis

And we can map this at the scale of the earth :
http://en.wikipedia.org/wiki/File:Seawifs_global_biosphere.jpg
Source :
http://oceancolor.gsfc.nasa.gov/SeaWiFS/BACKGROUND/Gallery/index.html

Good night :)

[Nathan Craig]

Hi Jeff,

This is a really nice diagram, I envy your graphics skills (mine are somewhat lacking). My primary feedback on the diagram is more of a comment of the thought experiment kind than an actual critique of any real significance. Visualizing NIR, like UV, is inherently difficult because it is invisible. In the diagram, NIR is represented as violet. However, NIR has a longer wavelength than red while violet is one of the shorter wavelengths of the visible spectrum. I wonder if confusion would arise by representing NIR with a color (violet) that actually sits at the short end of the visible spectrum. Two alternatives for showing NIR that came to my mind are either orange, which is also shorter than red, or a dashed outline. To me the dashes are, in some way, a better expression of the physics. The dashes would define the limits of the band, but by lacking color it might help express the fact that our eyes can't see reflected energy in these long wavelengths. Unfortunately, it might not look good in the diagram. I don't know, but these were just some thoughts. Besides, in the end white is just an equal mixture of RGB so in many ways the issue can't be escaped.

Also, if anyone is interested in delving further into absorption bands, the Jet Propulsion Laboratory (JPL) has a really good spectral library that is free to download. It could be very useful when trying to engineer cameras to target specific materials that are in spills, waste, runoff, etc.

http://speclib.jpl.nasa.gov/

Drawing on this library, it is possible to envision other kinds of band ratios, like ones that target salinity. Here is an example that I am working on with ASTER data, which by the way can also be obtained for free.
http://www.personal.psu.edu/nmc15/blogs/anthspace/2009/05/aster-band-ratios-for-soil-salinity.html

Field tests to validate the remote sensing can be performed with simple and inexpensive salinity and conductivity meters.

Regards,
Nathan

--
Nathan Craig, Ph.D., RPA
Assistant Professor
Department of Anthropology
409 Carpenter Building
Pennsylvania State University
University Park, PA 16802
Phone: 814.863.4388
FAX: 814.863.1474
www.anthro.psu.edu
andes.la.psu.edu
works.bepress.com/craig
pennstate.academia.edu/NathanCraig

[Jeffrey Warren]

great feedback nathan, esp. about the violet color. I'd actually chosen violet because that's the color the near infrared shows through our simple film filters, like in this image: http://publiclaboratory.org/notes/warren/1-7-2011/great-shot-ndvi

this is because the IR goes through all three of the RGB filters, but presumably more through the red and blue channels than the green one -- basically a coincidence, but I thought the visual correspondence would help people keep them straight. 

however, when you view it through a spectrometer, near infrared looks orange. Dunno why that's the case; maybe it's just that the cheap IR filter we're using actually leaks R and B, not that 'real' infrared goes through those filters more easily. Example:

http://publiclaboratory.org/notes/warren/12-24-2010/very-clear-spectrum-through-pineapple

Maybe if I used purple but then did some effect on it to indicate that it was invisible -- dashes, dots, etc.

I like the solution of orange since it works for thinking about spectrometers. 

jeff

[Nathan Craig]

I really like the pineapple image! I see your points about the how the NIR looks violet in the camera and orange in the spectrometer. That is very interesting.

Regards,
Nathan

[Adam Griffith]

Nice message string, all, and nice graphic, Jeff.

I like it, but I think of this process being dominated by the absorption of the red and blue light by the plant leaves.  Maybe if you somehow convey that the energy in the photons from the blue light and red light are actually captured in the leaves?  I don't know how to do this.  If you exaggerate the width of the incoming light bars and drastically reduce the width of the reflected red and blue light bars, that might help.  Or you could eliminate the red and blue bars being reflected off the leaf entirely.  Here's why: chlorophyll a and chlorophyll b and carotenoids actually absorb over 50% of light at certain wavelengths.

chlorophyll a peak absorption is 430 and 662 nm
chlorophyll b peak absorption is 453 and 642 nm
carotenoids peak absorption is 460-550 nm

http://www2.mcdaniel.edu/Biology/botf99/photo/p3igments.html

Your selection of Trillium sp. for the plant is great too!

Adam

[Adam Griffith]

OK, so thinking about this even further....

When asking why plants reflect infrared, we need to remember that they reflect EVERYTHING except what they absorb.  Right?

Adam

[Jeffrey Warren]

Cool. I revised the plants diagram and added one to show how we use different filters to get both visible and infrared photos:

http://publiclaboratory.org/tool/near-infrared-camera

Then I added an annotated version of the NDVI and NRG images to walk people through how to interpret them. Though NDVI interpretation can be nuanced (grass vs. trees, etc) I want to get people comfortable with reading them. Adam & Nathan - what do you think about the labels?

http://publiclaboratory.org/wiki/ndvi

One thing I'm interested is doing these kinds of diagrams right on the web page -- with an 'Edit' button. That way you could edit the diagrams, move labels and images, change colors and text, right in the browser. This is a bit tangential to Public Laboratory but i just think it would be kind of amazing to collaborate on images online in a 'graphical wiki' interface. It would also be great for annotating and collaboratively interpreting image data.

Jeff

[Adam Griffith]

Jeff, those figures and graphics are excellent.  Good labels.  They are well put together and convey the complexity of what is happening efficiently.  I echo Nathan's envy of your graphics 

Did Nathan get your last message?  He wasn't a recipient that I saw.

Adam D. Griffith

Western Carolina University

[Nathan Craig]

@ Adam: Thanks for ensuring I'm included; I'm now on the publiclaboratory list so I'll get messages posted to the group. By the way, I looked at your faculty page and found that you are doing super interesting stuff!

@Jeffrey: The new graphics look great, I really like them. By the way, I found the publiclaboratory page easy to edit; I did some minor grammar and spelling edits.

Warm regards,
Nathan

[Jeffrey Warren]

great to hear, thanks all.

[Adam Griffith]

@Nathan - same to you regarding your interests!

Have you tried this approach?  http://palentier.blogspot.com/2010/12/how-to-create-digital-elevation-model.html

Adam

[Nathan Craig]

@Adam: Some of the pipeline that Mark Willis used builds off of earlier attempts that I made to use open source software to transform structure from motion point clouds to real world coordinates. You'll see that Mark mentions this work down around step 15.

Mark Willis is super smart. His approach seems like it would work really well and strikes me as a significant streamlining of my earlier attempts.

More recently, I have been using Agisoft PhotoScan. It is not open source, but it is cheap and very powerful. I find it quite useful for generating photo textured 3D shapes from kite photographs. Using the PSU super computer facilities, I've been able to solve projects with as many as 500 images. This is not a good approach for the open source community, but it has been an interesting adventure to see how far PhotoScan can go...with sufficient RAM quite far it seems. When working with under 100 photos, a "normal" 64-bit workstation can do the job.

Other structure from motion software includes bundler and photocity. However, in conversations with Greg Downing, he presented a rather convincing case that PhotoSynth is actually faster than Bundler. As I have turned to PhotoScan, I don't use either Bundler or PhotoSynth that much these days. PhotoScan really likes near ortho images with lots of overlap; too much pivot in the Z (up and down) axis and PhotoScan doesn't like it. I use a gyro stabilized KAP rig, or when the wind is bad I tape the camera right to the picavet and hope for the best. As long as I can keep the camera pointed straight down, I get pretty consistently good results from photoscan. I think it would work very well for BAP. Nice thing about photoscan and the near ortho photographs is that the same images can also be used for making large planar panoramas--like the awesome ones regularly shown here.

By the way, here is a planar panorama I generated using the gyro stabilized rig and Kolor Autopano Giga. The image is a half a gigapixel and it is composed of about 300 images. AutopanoGigia isn't open source and costs about 300$, but it does a nice job with planar panoramas--at least on relatively flat subjects.

http://www.gigapan.org/gigapans/c768bab4245a5671a0b1db569e9668c7/

Regards,
Nathan

[ashele...@gmail.com]

I want to know about your diagram. What analysis did you take for your reflectance diagram?

[Jeffrey Warren]

Hi! At the time, I was reading all over the place about infrared plant analysis; some of this ended up on this wiki page: https://publiclab.org/wiki/ndvi

If you're interested, I've been "tidying" and updating a lot of our site content on infrared plant analysis on this page recently, including background and recent challenges: https://publiclab.org/wiki/infragram

(also CCing in the infrared discussion list)

Thanks!

J

--
Post to this group at publiclaboratory@googlegroups.com

Public Lab mailing lists (http://publiclab.org/lists) are great for discussion, but to get attribution, open source your work, and make it easy for others to find and cite your contributions, please publish your work at http://publiclab.org
---
You received this message because you are subscribed to the Google Groups "The Public Laboratory for Open Technology and Science" group.
To unsubscribe from this group and stop receiving emails from it, send an email to publiclaboratory+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.