1. Select some features from the hyperspectral image like calculate gradient

2. With that value make a feature vector for the positive and negative images

3. store that feature vector into a file .csv or .dat format

4. select the classification algorithm like SVM or something else

5. make a model with the training set

6. test that model

1. Select some features from the hyperspectral image like calculate gradient

2. With that value make a feature vector for the positive and negative images

3. store that feature vector into a file .csv or .dat format

4. select the classification algorithm like SVM or something else

5. make a model with the training set

6. test that model

For use SVM or Support Vector Machine you need training matrix and group matrix

note : Svm support two group , e.x group 0 and group 1

First : calculate some feature for training matrix

Second : Design matrix group

Third : after step 1 and 2 you tarining svm how ?

svmstruct=svmtrain(training,group, ...

'Autoscale',true, 'Showplot',false, 'Method','QP', ...

'BoxConstraint',2e-1, 'Kernel_Function','rbf', 'RBF_Sigma',1);

forth : after step third you test your algorithm how ?

GROUP=svmclassify(svmstruct,sam,'Showplot','false');

if GROUP==0

disp('***************');

else

disp('***************');

end

Please suggest some feature for spatial and spectral classification