Table 9 Performance of different ANN models for prediction of measured parameters and SY of dry bean using RGBIs and SRIs together.

From: Water status and plant traits of dry bean assessment using integrated spectral reflectance and RGB image indices with artificial intelligence

Model

Optimal SRI and RGBIs

Hyper-parameter

(Z, L, N, I)

Training

Testing

R 2

MSE

R 2

MSE

ANNB-WB3

ExG, CIVE, TGI, IPCA, \(\:{\text{S}\text{R}\text{I}}_{\text{580,1130}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{586,1130}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{636,630}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{642,632}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{648,622}}\)

(Logistic, 2, 5, 700)

0.97***

2.09

0.96***

2.35

ANNB-DB3

ExG, CIVE, COM, TGI, IPCA, MExG, \(\:{\text{S}\text{R}\text{I}}_{\text{580,1130}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{586,1130}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{636,630}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{642,632}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{648,622}}\)

(ReLU, 1, 9, 700)

0.99***

0.10

0.987***

0.20

ANNB -CMC3

RGVBI, ExG, CIVE, TGI, IPCA, MExG, \(\:{\text{S}\text{R}\text{I}}_{\text{586,1130}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{636,630}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{642,632}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{648,622}}\), \(\:{\text{S}\text{R}\text{I}}_{662,\:610}\)

(Logistic, 1, 8, 500)

0.93***

0.41

0.88***

0.53

ANNB-SPAD3

ExG, CIVE, COM, TGI, MExG, \(\:{\text{S}\text{R}\text{I}}_{\text{636,630}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{642,632}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{648,622}}\), \(\:{\text{N}\text{D}\text{I}}_{\text{686,620}}\)

(ReLU, 2, 5, 500)

0.935***

0.61

0.933***

0.71

ANNB-SWC3

ExG, CIVE, COM, TGI, GLI, MExG, \(\:{\text{S}\text{R}\text{I}}_{\text{574,1134}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{580,1130}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{586,1130}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{636,630}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{642,632}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{1104,710}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{1120,1142}}\)

(Tanh, 1, 8, 800)

0.98***

0.77

0.977***

0.81

ANNB-SY3

ExG, ExGR, CIVE, COM, TGI, MExG, \(\:{\text{S}\text{R}\text{I}}_{\text{574,1134}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{580,1130}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{586,1130}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{636,630}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{642,632}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{1104,710}}\), \(\:{\text{S}\text{R}\text{I}}_{\text{1120,1142}}\)

(Tanh, 1, 6, 500)

0.99***

0.39

0.988***

0.57

  1. Z, L, N, and I represent activation function, layers number, neurons number in each layer, and iterations number, respectively.
Back to article page