Public Lab Research note


NDVI difficulties with infrablue filter

by warren | June 02, 2013 14:55 | 89 views | 4 comments | #8018 | 89 views | 4 comments | #8018 02 Jun 14:55

Read more: publiclab.org/n/8018


Some folks at the pirateship and I used a Roscoe 2008 filter which had not been very well white balanced to take some aerial photos of our building and neighborhood, but I found that the resulting NDVI images (made with Ned Horning's Photo Monitoring plugin) didn't work very well -- the trees and lawn did not display with high NDVI values. What's the problem? Can I do a better post-white-balancing to recover useful data?

Was the Roscoe 2008 not as good as the 2007? I notice that Chris F did not have one in his exhaustive comparative search for filters, in this research note. But the filter looks visibly similar... what do you think? I can do a spectrometer scan of both for comparison.

IMG_0630.JPG

ndvi-5.jpg


4 Comments

Jeff, do you have an image of the lut you used? You can open a lut in Fiji by file/open.

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

Reply to this comment...


The Rosco 2008 should be one of the best filters for making NDVI because it blocks all the red light better than others. It has only 40% of the overall transmission of the 2007, so exposures will have to compensate by more than one f/stop. It blocks a lot of the green light too, so false color Infrared (NBG) might be harder.

The first thing I would do is to make a floating point NDVI image, open it in Fiji, and apply the lut of choice to it. That might look different from the color NDVI produced by the plugin.

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

Reply to this comment...


Can I do a better post-white-balancing to recover useful data?

If you adjust the white balance on the infrablue photo, or make any other changes to the color balance, you will get different NDVI values. By trial and error you can get NDVI values that make more sense. Whether the new NDVI image will be useful data is less clear. I guess you could start by adjusting the color so the photo looks more yellow like other infrablue photos. Maybe that will make the NDVI image look more meaningful.

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

Reply to this comment...


I have a feeling it is a white balance problem. The blue channel seems to be off. I would expect vegetation to be much darker in the blue channel. The green and red channels look very similar.

Reply to this comment...


Login to comment.

Public Lab is open for anyone and will always be free. By signing up you'll join a diverse group of community researchers and tap into a lot of grassroots expertise.

Sign up