Table 7.

Precision, recall, F1-scores of five F-CNNs models, and two baseline methods for detecting crop field and boundary from three channels (blue-green-red) high-resolution satellite images.

	Boundary			Cropland
	Precision	Recall	F1-Score	Precision	Recall	F1-Score
	Training
Random Forest	0.82	0.63	0.71	0.94	0.98	0.96
FCN-DKConv6	0.85	0.85	0.85	0.96	0.97	0.96
U-Net	0.84	0.85	0.84	0.97	0.97	0.97
SegNet	0.84	0.84	0.84	0.96	0.96	0.96
DenseNet56	0.93	0.93	0.93	0.98	0.98	0.98
DenseNet67	0.93	0.94	0.93	0.99	0.99	0.99
DenseNet103	0.94	0.95	0.94	0.99	0.99	0.99
Test
Random Forest	0.48	0.21	0.29	0.87	0.96	0.91
FCN-DKConv6	0.56	0.56	0.56	0.92	0.93	0.92
U-Net	0.64	0.65	0.65	0.93	0.93	0.93
SegNet	0.67	0.66	0.66	0.93	0.94	0.94
DenseNet56	0.74	0.74	0.74	0.94	0.95	0.94
DenseNet67	0.77	0.75	0.76	0.95	0.95	0.95
DenseNet103	0.78	0.76	0.77	0.95	0.96	0.96