I have tasseled cap component of Brightness (TBC) and i want to calculate its accuracy with NDVI in ENI using Confusion matrix. How do i find those values from NDVI that are comparable to TBC.
Although, i have used OTSU in MATLAB on NDVI, but i'm confused as NDVI mainly defines vegetation.
Your question is unclear.
Some suggestions:
Sure, NDVI was desighned as vegetation index. TCT Brightness correspond to land surface as a whole. It is completely different feature and used in some cases as a companion index in addition to other TCT features. If you mean TCT Greenness, you should use correlation coefficients for certain samples rather than confusion matrix.
As for TCT brightness itself, you may compare it with,say, panchromatic band of the same satellite sensor if it provide such band.
But I'm confused by the reason...
Maybe the article attached can help.
Good luck!
Maria
@ Maria...
I already did comparison between NDVI and T-cap Greenness, with 70% accuracy.
I'm also confused at the same thing that how can i compare T-cap Brightness with NDVI?
And by you suggestion of comparing with PAN band, why is this so?
Dear Atiqa,
I attach a table from well known 'classical' article issued by E.Christ & R.C. Cicone. A Physically-Based Transformation of Thematic Mapper Data-The TM Tasseled Cap // IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. GE-22, NO. 3, MAY 1984
As you surely know, Brightness feature is a weighted sum of all VIS, NIR and SWIR band values (based on soil line). Brightness coefficients are underlined by red. All of them are positive. PAN band directly summarize VIS brightness, so you can compare it to find out if there are some valuable reasons to calculate TCT Brightness. This correspond to land cover as a whole.
Greenness feature correspond to vegetation as it provide highest positive loading in Red band and negative one in NIR band (underlined by green). Anyway Greenness differ from simple NDVI index. NDVI provide a contrast between Red and NIR bands (also correspond to vegetation), as Greenness is a weighted sum of all band values (like Brightness) BUT the weights (TCT coefficients) are very different being positive in VIS and mostly in RED band and negative in NIR band!
Such a way, NDVI and Greenness usually correlate relatively well, but Brightness is very different from NDVI as high Brightness values may be caused by soil brightness or build-up or so.
So what u suggest ... skip NDVI and use another index for comparison?
- Badges
- Science topic
- Similar topics
- Analytical Chemistry
- Extraction