This was my undergraduate final year thesis project. This project basically classifies hyperspectral images usning a reduced version of Kernel PCA.
USAGE:
In RKPCA.m file, load the hyperspectral image in .mat format. Select test and train data from the resultant image using multispec software. Put the test and train data in DataSet folder as KPCAtest.txt and KPCAtrain.txt file. From SVM.m file, load the train and test data Finally wait for classification accuracy result. Output time depends on the train and test data dimensionality