Orthogonal design based genetic algorithm for spatial resolution enhancement of hyperspectral images

In this work, the use of orthogonal design based genetic algorithm for enhancing spatial resolution of hyperspectral images is proposed. High spectral resolution of hyperspectral images allows to extraction of information at sub-pixel level. For this purpose, firstly an endmember extraction operation is performed by using Vertex Component Analysis (VCA) and then abundances for each endmember is calculated by Fully Constrained Least Square Algorithm (FCLS). The spatial resolution of hyperspectral images aimed to enhanced by using endmembers and abundances. The proposed method is tested for synthetic hyperspectral images that known ground truth map and experimental results show that orthogonal design based genetic algorithm improves the performance.