Skip to main content

Table 3 Coefficients of determination (R2) and root mean square error (RMSE in mg·dm−2) of the indices developed for phenotyping EW in leaves

From: High-resolution spectral information enables phenotyping of leaf epicuticular wax in wheat

Index

Model parameters

a

B

R2

RMSE

95% CI

p-value

Broad indices / RGB and NIR spectral bands

 EWI-1 Blue/Red

0.213

0.04

0.44

1.19

1.037–2.17

 < 0.0001

 EWI-2 Blue/NIR

0.07

0.13

0.39

1.18

0.98–1.98

 < 0.0001

 EWI-3 (NIR-Red)/Blue

− 0.93

− 0.01

0.31

1.19

1.04–1.97

 < 0.0001

 EWI-4 (Red2-Blue)/(Red-Blue2)

− 0.09

− 0.03

0.32

1.19

1.09–2.55

 < 0.0001

Narrow indices / two narrow spectral bands

 EWI-5 676

0.019

0.005

0.45

0.97

0.75–1.21

 < 0.0001

 EWI-6 658/712

0.12

0.03

0.52

1.02

0.70–1.36

 < 0.0001

 EWI-7 625/706

0.22

0.05

0.50

0.96

0.67–1.28

 < 0.0001

 EWI-8 694/625

− 0.006

− 0.002

0.42

1.08

0.96–1.55

 < 0.0001

 EWI-9 (670–718) / (670+718)

− 0.85

0.03

0.51

1.04

0.61–1.54

 < 0.0001

 EWI-10 (691–661) / (691+661)2

4.92

− 1.03

0.48

0.99

0.74–1.27

 < 0.0001

 EWI-11 (1/661)—(1/694)

29.13

− 5.73

0.48

1.01

0.71–1.35

 < 0.0001

 EWI-12 (622/718)-1

0.62

0.12

0.51

0.99

0.74–1.28

 < 0.0001

Narrow indices / three narrow spectral bands

 EWI-13 625 (1/736 – 1/832)

0.008

0.004

0.65

1.01

0.622–1.426

 < 0.0001

 EWI-14 (625–736) / 832

0.02

0.007

0.62

0.98

0.65–1.35

 < 0.0001

  1. R2 was calculated in the training set by a leaving one out cross validation analysis (LOOCV) and the RMSE was estimated in the validation set. B is the slope of the line and a is the intercept of the dependent variable