Conformal geometric algebra based band selection and classification for hyperspectral imagery

Conformal geometric algebra (CGA) has several advantages such as consistent geometric representation, compact algebra formulae, efficient geometric computing, coordinate free, and dimensionality independent etc., it can provides a new mathematical tool for hyperspectral dimensionality reduction. In this paper, an efficient band selection and classification approach for hyperspectral imagery based on CGA is proposed. In order to achieve more concise, fast, robust hyperspectral dimensionality reduction, the CGA-supported band selection method in conformal space is designed. The experiment results show that the CGA-based band selection algorithm outperforms the popular sequential forward selection (SFS) and particle swarm optimization (PSO) with lower cost for hyperspectral band selection.