Hyperspectral Image Classification by AdaBoost with Decision Stumps Based on Composed Feature Variables

Over the past few decades, a considerable number of studies have been made on statistical classification methods for hyperspectral imagery. For classification of hyperspectral data, we must take care of a curse of dimension and computation cost. For the problem, we propose AdaBoost by decision stumps based on composed feature variables. We show that the method can be processed in acceptable time for AVIRIS data. The proposed method obtains a more accurate result compared to kernel based NN and SVM. We also assess features of hyperspectral data from the obtained classifiers. The proposed method can imply the relative importance of the feature for classification.