Accessing Hyperspectral Images Using MATLAB

I am doing a project on image fusion of Hyperspectral Images

It involves calculating the local variance for each band .

I downloaded Hyperspectral Images from this website --> http://personalpages.manchester.ac.uk/staff/david.foster/Hyperspectral_images_of_natural_scenes_02.html

But the images are in .mat format, and I am not able to access the data through Wavelet Toolbox .

What should I do so as to separate the image into various bands ?

How do i find the local variance for each image ?

Answer 1

[Brett Shoelson]

If the images are stored in .mat files, you should be able to LOAD them with core MATLAB--no special Wavelet-reading functions needed.

Then it's a matter of indexing. Assuming your spectral cube A is m x n x p (with p spectral bands), you would pick out the first band with A(:,:,1), and the pth band with A(:,:,p). (Et cetera.)

Brett

[S@M]

• %% Simple Load Hyperspectral data and Crossponding Ground Truths.
• load('Indian_pines_corrected.mat');
• load('Indian_pines_gt.mat');
• img = indian_pines_corrected;
• gt = indian_pines_gt; clear indian*
• %% Display individual Band.
• imagesc(img(:,:,i)); %% i could be any depending upon your choice raning from 1-224 in this case.
• imagesc(gt); %% Show Ground Truths.

• %% Hope this helps.
• %% Don't forget to read the related works.
• %% A New Statistical Approach for Band Clustering and Band Selection Using K-Means Clustering.
• %% AIK Method for Band Clustering Using Statistics of Correlation and Dispersion Matrix.
• %% Hyperspectral unmixing using statistics of Q function.
• %% unmixing and target detection of hyperspectral imagery using OSP.
• %% Metric similarity regularizer to enhance pixel similarity performance for hyperspectral unmixing.
• %% Unsupervised geometrical feature learning from hyperspectral data.
• %% Graph-based Spatial-Spectral Feature Learning for Hyperspectral Image Classification.
• %% Fuzziness-based active learning framework to enhance hyperspectral image classification performance for discriminative and generative classifiers.