A Research on the Combination Strategies of Multiple Features for Hyperspectral Remote Sensing Image Classification

Table 8

Classification results by different features for PU dataset, including OA, kappa index, and execution time.

PU		SPEC	GF	TF	GF + TF	TF + GF + SPEC
Direct combine	OA (%)	83.61	85	92.78	98.6	98.48
	Kappa	0.775	0.797	0.904	0.982	0.980
	Time (s)	8.91	4.36	2.38	6.6	8.5
LE framework	OA (%)	80.89	83.66	91.78	98.44	97.69
	Kappa	0.741	0.778	0.89	0.979	0.969
	Best	25	15	40	55	45
	Time (s)	9.23	5.48	6.62	7.91	8.48
Multiple kernel	OA (%)	83.61	84.91	96.11	98.08	97.86
	Kappa	0.775	0.796	0.948	0.975	0.972
	Time (s)	8.91	4.76	5.96	7.03	12.05