Combined Kernel-Based BDT-SMO Classification of Hyperspectral Fused Images

Table 7

Comparison of BDT-SMO classification approaches under different kernels (zone 2).

Kernels	Overall accuracy (%)	Average accuracy (%)	Kappa coefficient	Training time (s)	Test time (s)	Standard deviation	Optimized parameters
Linear	80.11	79.33	0.7841	43.15	16.41	1.85
Polynomial	81.19	80.06	0.7859	43.59	22.79	1.75	,
Sigmoid	82.49	81.39	0.8009	44.95	15.29	1.61	,
SRBF	84.05	83.37	0.8254	42.11	14.59	0.59	,
CRBF	86.15	85.52	0.8483	43.14	13.95	0.62	,; ,