A Quantum Annealer for Subset Feature Selection and the Classification of Hyperspectral Images

Hyperspectral Earth-observation images contain many spectral bands, some of which may be redundant or noisy. This project uses quantum annealing both to select highly informative spectral-band subsets and to train quantum classifiers for hyperspectral-image classification. The workflow is tested on the AVIRIS Indian Pines dataset, with the feature-selection and classification tasks formulated as QUBO problems for a D-Wave annealer. The study demonstrates classification quality comparable to conventional approaches while highlighting the practical requirements of real-world datasets.