Spectral-Spatial Joint Classification of Hyperspectral Image Based on Broad Learning System

At present many researchers pay attention to a combination of spectral features and spatial enhance hyperspectral image (HSI) classification accuracy. However, the in some methods are utilized insufficiently. In order further improve performance HSI classification, spectral-spatial joint based on broad learning system (BLS) (SSBLS) method was proposed this paper; it consists three parts. Firstly, Gaussian filter is adopted smooth each band original spectra information remove noise. Secondly, test sample’s labels can be obtained using optimal BLS model trained with smoothed by filter. last, guided performed correct results contextual for improving Experiment real datasets demonstrate that mean overall accuracies (OAs) ten experiments 99.83% Indian Pines dataset, 99.96% Salinas 99.49% Pavia University dataset. Compared other methods, paper has best performance.