Stavroula Giannakopoulou
New Member
Posts:49  
| 04 Feb 2020 05:49 PM |
|
|
Hello everyone,
I have a question regarding the routine of Spectral Hourglass Wizard. I want to implement the procedure of Wizard at my hyperspectral data. So my question is if I should try the MNF procedure. My thought is to see if the results without the step of MNF are also valid. Thank you very much for your help!!!
|
|
|
|
MariM
Veteran Member
Posts:2396 
| 05 Feb 2020 12:03 PM |
|
|
You can exclude the MNF transform if you like but it was designed to reduce dimensionality and redundancy in highly correlation hyperspectral imagery and eventually be used in linear spectral unmixing or MTMF. Some routines, such as MTMF, require the spectra to be in MNF space so it will depend on the end use of the data.
|
|
|
|
Stavroula Giannakopoulou
New Member
Posts:49
| 05 Feb 2020 04:04 PM |
|
|
OK! I see what you mean, but I was thinking of implementing it because I want to have a result only for the SAM method! Thank you very much!!!
|
|
|
|
MariM
Veteran Member
Posts:2396
| 06 Feb 2020 05:34 AM |
|
|
SAM would not require MNF but it is a 'whole pixel' classifier. LSU and MTMF is designed to perform sub-pixel analysis, so it is much more important to find the true endmembers in the image to find the sub-pixel material abundances.
|
|
|
|