Implement MNF transform

[Agustin Lobo]

I miss a MNF transform for increasing SNR, particularly useful in cases there is a uniform subset

or image in which calculate the noise covariance matrix (i.e., field or lab images acquired with hyperspectral imaging spectrometers).

https://www.l3harrisgeospatial.com/docs/MinimumNoiseFractionTransform.html

[Agustin Lobo]

Note MNF is not the same as NMF (AFAIK). It seems MNF is not implemented in scikitlearn, but it is in

https://pysptools.sourceforge.io/noise.html#minimum-noise-fraction-mnf-transformation

and

http://www.spectralpython.net/

MNF is aka noise-adjusted PCA

[Andreas Janz]

Thanks for the PySptools tip. They offer some interesting hyperspectral functions (MNF, SAM, linear unmixing, etc.) widely known from good old ENVI Classic.

I will have a look at it and hopefully connect MNF and other routines to the EnMAP-Box.

[Andreas Janz]

Removing version: 3.10 (automated comment)

[Agustin Lobo]

Any prospective date for this issue to be implemented?

‌

[Andreas Janz]

No , not so far.