Table 4. Classification results of PLS-DA and RC-PLS-DA models.

Models	Type	Number of variables	Calibration			Prediction
			No.a	Correct	CA/%b	No. a	Correct	CA/% b
PLS-DA	Intact	396	34	34	100	17	16	94.12
	Cracked	396	34	34	100	17	17	100
	All	396	68	68	100	34	33	97.06
RC-PLS-DA	Intact	18	34	31	91.18	17	17	100
	Cracked	18	34	34	100	17	13	76.47
	All	18	68	65	95.59	34	30	88.24

a: Number of samples; b: Classification accuracy.