Why is my Green Chlorophyll Index (GLI) scale so large?

669
6
04-01-2019 02:58 PM
New Contributor II

Hi everyone,

I am trying to calculate my green chlorophyll index (GCI), one of my many vegetation indices that I'm trying to calculate. I am working with imagery from a Micasense RedEdge 3 camera (with proper radiometric calibration procedures already completed). I thought that you are supposed to get data values between -1 and 1, but instead, my values range from 0.09-14.4, as you can see by the image that I have attached. 

The equation for the GCI, as defined by the literature, is: (NIR/Green) - 1

My dataset bands are defined here, for this example: Green = "band_2" and NIR = "band_5"

The equation I used for Raster Calculator is this: Float("band_5") / Float("band_2") - 1

Any help would be greatly appreciated! Thanks guys!

(P.S. This is a picture of a grain sorghum research plot)

Reply
0 Kudos
6 Replies
MVP Esteemed Contributor

That doesn't look like a normalized index.

If band 5 has a value of 100 and band 2 has one of 50... ie raw pixel values, then you won't get a number that falls in that range.  As the band values supposed to be 'raw' or normalized?

Imagery and Remote Sensing m‌ight be a better place to move this since it really isn't an ArcGIS Pro issue

Esri Regular Contributor

Agreed with Dan

https://www.indexdatabase.de/db/i-single.php?id=128 confirms you have the correct equation for GCI but it's clearly not a normalized index.  If NIR=250 and green = 1, GCI will be 249

-1 < GCI < 254, and you'll also have to check for Green = 0 to avoid division by zero.

New Contributor II

Thank you both for your inputs. I appreciate the help a lot!

Thanks for directing me to the Remote Sensing option Dan, I appreciate that!

Reply
0 Kudos
New Contributor II

Hi Issac,

Just out of curiosity, did you get chance to normalize the data?. I have also similar values and don't know how to normalize the data. Thanks

Reply
0 Kudos
New Contributor II

Hi Kushendra,

I did not get a chance to normalize this data. Since it is not a normalized dataset, I am thinking to just use the data as is.

If you absolutely have to normalize it, I would try to extract the data (i.e. using the "zonal statistics as table" tool) and then normalize it in this method: https://stats.stackexchange.com/questions/70801/how-to-normalize-data-to-0-1-range

Not claiming to be an expert, but this might help!

New Contributor II

Thanks Isaac!

Reply
0 Kudos