plot the reflectance spectra of three different pixels for all bands in a hyperspectral image

I need to develop a program that generates a binary mask of a hyperspectral image using spectral band difference based segmentation, and then select three different pixels in the image to plot the reflectance spectra for all the bands. There are 244 individual bands, so each image is considered a separate band. I attached 4 different bands for reference. The goal would be to select a given pixel on the pot, plant, and background and compare the reflectance spectra.

[Walter Roberson]

Once you have created a 320 x 455 x 244 data array from the input data, then for any given pixel hyperdata(ROW, COL, :) then how would you calculate the reflectance spectra ?

[Mason Lien]

Creating the data array for the input data is something I am curretnly trying to complete but having issues. I found this for loop to stack the bands but am having issues getting the loop to work. Not sure if this is the best approach. I'm getting a read permission error.

As for calculating the reflectance spectra, I'm not sure if I need to calculate, but give a general observation of how the 3 selected pixels are plotted. I may be mistaken.

P = 'absolute/relative path to where the files are saved';
N = 244;
C = cell(1,N);
for k = 1:N
    F = sprintf('%_0_0.png',k);
    C{K} = imread(fullfile(P,F));
end
A = cat(4,C{:})

[Walter Roberson]

That reading code looks plausible -- though I suspect cat 3 might be more appropriate (your png are likely grayscale.)

[Mason Lien]

Following up - before I can create the array these images would need to be converted to grayscale correct?

[Walter Roberson]

I just checked and your images are represented as RGB but have only grayscale content. So I suggest

P = 'absolute/relative path to where the files are saved';
N = 244;
C = cell(1,N);
for k = 1:N
    F = sprintf('%_0_0.png',k);
    rgb = imread(fullfile(P,F));
    C{K} = rgb(:,:,1);    %any colorpane does for grayscale
end
A = cat(3, C{:})

[Subhadeep Koley]

@ Mason Lien For visualization of the spectra / hyperspectral image, you can use the hyperspetralViewer App, which comes under Image Processing Toolbox's Hyperspectral Imaging Library support package, and can be downloaded from here.

[Mason Lien]

Walter and subhadeep, thanks for your respones! I was able to get what I needed.