Table 2.

The classification accuracies of RS-2 Flevoland (%).

Method	Water	Forest	Building	Farmland	OA	Kappa
Lenet-5 (Samples 1)	96.00	89.54	89.93	90.18	90.96	0.88
FFS-CNN (Samples 1)	97.90	91.91	93.99	94.98	94.40	0.92
Lenet-5 half (Samples 2)	93.42	80.44	84.15	92.72	87.34	0.81
FFS-CNN half (Samples 2)	94.01	87.22	90.04	94.47	91.26	0.86
Method in [39]	98.65	89.07	73.40	89.42	90.01	N/A