CODEWITHSUNDEEP
pythonpcasatellite-imagespectral-python
S.EB
S.EB

Reputation: 2226

How we can group the spectral bands into subgroups?

The simplest way is to group based on the neighboring wavelength groups in hyperspectral images, as shown in the following image, where there are $G$ different groups and the spectral bands are grouped in equally sized bands. For example, each group $$g_i$$ has 10 spectral neighboring bands grouped together.

enter image description here

I am looking for a more sophisticated technique to group the spectral bands. Moreover, as far as I know, this is not related to band selection since we need all bands. Do you know any research paper/study for this aim? I would appreciate it if you share the studies. Thanks

Upvotes: 0

Views: 103

Answers (1)

bogatron
bogatron

Reputation: 19179

There are multiple ways you could group bands but the first thing to answer is "Why are you grouping bands?"

If it is simply to break up a large file into manageable smaller files, then you would probably just break it up into groups of contiguous bands, as you have suggested.

You also state that this is not related to band selection since you need all bands but in fact, it might be. Even if you are keeping all bands, it might make sense to group bands by their utility or phenomenology. For example, you might want to group them by spectral range (e.g., VNIR, SWIR, MWIR, LWIR), since different spectral regions relate to different physical properties of the target of the image.

Upvotes: 1

Related Questions