A feature extraction method based on spectral segmentation and integration of hyperspectral images

Sayyed Hamed Alizadeh Moghaddam, Mehdi Mokhtarzade, Behnam Asghari Beirami

Abstract: Highlights•Pixel-based feature extraction methods are advantageous as they preserve the characteristics of spectral signature curves.•In contrast to the similar methods applying local search algorithms, the proposed method employs a global one known as PSO.•PSO can yield more accurate results as it can escape from local optimal solutions.•Experimental results proved the superiority of the proposed SSI over seven state-of-the-art methods in the literature.