TABLE 5.

The classification accuracy based on fusion data of AE and PCA features of different spectral sources.

Cultivar	Features	Variables	Model	Tr (%)	Val (%)	Te (%)	Model	Tr (%)	Val (%)	Te (%)	Model	Tr (%)	Val (%)	Te (%)
Zhefujing83	AE-PCA-HSI	46	SVM	98.33	100	98.08	LR	98.33	100	100	CNN	98.33	85.00	90.38
	AE-PCA-MIR	50		100	100	86.54		100	100	88.46		100	100	86.54
	AE-PCA-LIBS	178		100	90.00	61.54		99.17	95.00	67.31		100	95.00	73.08
	PCA-HSI-MIR	48		100	100	96.15		100	100	98.08		100	95.00	88.46
	PCA-HSI-LIBS	128		100	95.00	96.15		100	100	100		100	85.00	63.46
	PCA-MIR-LIBS	148		100	100	86.54		100	100	86.54		100	95.00	48.08
	PCA-HSI-MIR-LIBS	162		100	100	96.15		100	100	96.15		100	100	80.77
	AE-PCA-HSI-MIR	96		100	100	94.23		100	100	98.08		100	95.00	84.62
	AE-PCA-HSI-LIBS	224		100	95.00	75.00		100	95.00	75.00		100	95.00	75.00
	AE-PCA-MIR-LIBS	228		100	100	88.46		100	100	90.38		100	100	86.54
	AE-PCA-HSI-MIR-LIBS	274		100	100	92.31		100	100	92.31		100	100	78.85
AD516	AE-PCA-HSI	46		100	100	92.19		100	100	100		100	95.00	92.19
	AE -PCA-MIR	58		100	100	95.31		100	100	96.88		100	95.00	84.38
	AE -PCA-LIBS	167		100	85.00	79.69		100	100	84.38		100	95.00	73.44
	PCA-HSI-MIR	56		100	100	98.44		100	100	100		100	95.00	90.63
	PCA-HSI-LIBS	117		100	100	98.44		100	100	98.44		100	95.00	87.5
	PCA-MIR-LIBS	145		100	100	85.94		100	100	85.94		100	95.00	85.94
	PCA-HSI-MIR-LIBS	159		100	100	95.31		100	100	98.44		100	100	87.5
	AE-PCA-HSI-MIR	104		100	100	98.44		100	100	100		100	100	85.94
	AE-PCA-HSI-LIBS	213		100	100	82.81		100	100	90.63		100	95.00	73.44
	AE-PCA-MIR-LIBS	225		100	100	79.69		100	100	85.94		100	100	78.13
	AE-PCA-HSI-MIR-LIBS	271		100	100	82.81		100	100	100		100	90.00	78.13

Other abbreviations are similar to this fashion. Tr, training set; Val, validation set; Te, test set; AE-PCA-HSI, the concatenation of PCA features and AE features of HSI; PCA-HSI-MIR, the concatenation of PCA features of HSI and PCA features of MIR; AE-PCA-HSI-MIR, the concatenation of AE-PCA features of HSI and AE-PCA features of MIR.