I would like to determine endmembers using PPI method. The PPI is typically run on an MNF transform result.
However, I do not know when to apply an image fusion technique to the image data - is it before or after MNF transformation? Which technique can provide a better result!
Thanks for any help.
PPI is for endmember extraction, MNF is to remove noises and re-rank subspaces so that first dimension has the highest SNR.
So PPI after MNF.
I don't understand why there is an image fusion here.
Hi,
After reducing the data dimentionality by implementing MNF, PPI is typically applied to determine the pure pixels. So, there is no need of doing pan-sharpening on your dataset. N-dimentional Visualizer (in ENVI) should be applied afterwards, as an multidimensional analysis to identify the spectral endmembers.(usually those which are located at corners are selected!).
Hi and many thanks! I was in doubt about it! Now everything is ok again.
- Badges
- Science topic