Determination of Wheat Heading Stage Using Convolutional Neural Networks on Multispectral UAV Imaging Data

Table 3

Fitting effect of each band combination input fitting model.

Input band (nm)	Fitting model	Fitting results
		R^2	RMSEP
550 + 660 + 730 + 790 nm	NN	0.77	0.24
	SVR	0.78	0.29
	DT	0.83	0.25
	1D CNN + DT	0.95	0.24
550 + 660 + 730 nm	NN	0.14	0.12
	SVR	0.58	0.25
	DT	0.76	0.24
	1D CNN + DT	0.82	0.24
550 + 660 + 790 nm	NN	0.12	0.48
	SVR	0.68	0.27
	DT	0.74	0.23
	1D CNN + DT	0.92	0.29
550 + 730 + 790 nm	NN	0.11	0.58
	SVR	0.69	0.27
	DT	0.70	0.26
	1D CNN + DT	0.92	0.28
660 + 730 + 790 nm	NN	0.11	0.43
	SVR	0.68	0.37
	DT	0.74	0.36
	1D CNN + DT	0.76	0.42