Spectral Resampling from hyperspectral to multispectral
Show older comments
Can anyone help me to do this spectral resampling of hyperspectral image to multispectral image using matlab. Please anyone help me to do this.
3 Comments
Bjorn Gustavsson
on 2 Jun 2022
So you want to go from very many spectral wavelength-layers to a sensible number of them?
Nandita Sarkar
on 3 Jun 2022
Image Analyst
on 3 Jun 2022
OK, if youi don't like my Answer/solution below you can just scan every pixel and resample using interp1d.
Answers (2)
Image Analyst
on 2 Jun 2022
0 votes
How many wavelength bands are you starting with and how many do you think you want to end up with?
Do they need to be actual wavelength bands? If you want you could use pca (demo attached) and use the PC's as the new images.

Bjorn Gustavsson
on 4 Jun 2022
If you have a hyper-spectral image, I_hs, on the format [n_y x n_x x n_lambda] and you want to reduce the number of wavelength-channels you might do something like this (hoofed live-edited and no-tested):
function I_ms = hyper2multi(I_hs,w_lambda)
sz_Ihs = size(I_hs);
sz_w = size(w_lambda);
I_ms = zeros([sz_Ihs(1:2),sz_w(2)]);
for i_lout = 1:sz_w(2)
for i_lin = 1:sz_Ihs(3)
I_ms(:,:,i_lout) = I_ms(:,:,i_lout) + I_hs(:,:,i_lin)*w_lambda(i_lin,i_lout);
end
end
end
where w_lambda should be an [n_lambda x n_multi] array with weighting-factors for the spectral response you want in your multi-spectral image (all zeros except a single 1 in a column if you want to extract a single wavelength-channel)
HTH
Categories
Find more on Hyperspectral Image Processing in Help Center and File Exchange
Products
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!