A nonlinear regression classification algorithm with small sample set for hyperspectral image

摘要

A column generation kernel technology based nonlinear regression classification method for hyperspectral image is proposed in this paper. The nonlinear extension for the collaborative representation regression is utilized in the joint collaboration model framework. The proposed algorithm is tested on two hyperspectral images. Experimental results suggest that the proposed nonlinear algorithm shows superior performance over other linear regression-based algorithms and the classical hyperspectral classifier SVM.