Key Parameters for Maize Leaf Moisture Measurement Using NIR Camera With Filters Based on Hyperspectral Data

Peng-cheng LI; Han LIU; Long-lian ZHAO; Jun-hui LI

doi:10.3964/j.issn.1000-0593(2021)10-3184-05

Journals >Spectroscopy and Spectral Analysis >Volume 41 >Issue 10 >Page 3184 > Article

Spectroscopy and Spectral Analysis
Vol. 41, Issue 10, 3184 (2021)

Key Parameters for Maize Leaf Moisture Measurement Using NIR Camera With Filters Based on Hyperspectral Data

Peng-cheng LI^*, Han LIU^{1; 2;}, Long-lian ZHAO^{1; 2;}, and Jun-hui LI^{1; 2; *;}

Author Affiliations

1. College of Information and Electrical Engineering, China Agricultural University, Beijing 100083, China

show less

DOI: 10.3964/j.issn.1000-0593(2021)10-3184-05 Cite this Article

Peng-cheng LI, Han LIU, Long-lian ZHAO, Jun-hui LI. Key Parameters for Maize Leaf Moisture Measurement Using NIR Camera With Filters Based on Hyperspectral Data[J]. Spectroscopy and Spectral Analysis, 2021, 41(10): 3184 Copy Citation Text

show less

Fig. 1. Scatter plots of moisture reference and predicted values for calibration set (a) and prediction set (b) with 25 nm bandwidth filter

Download full size | View in the Article

Model	Feature wavelength	R²	RMSE
1 wavelength	1 151	0.935	1.374
2 wavelengths	1 151, 1 403	0.962	1.074
3 wavelengths	1 151, 1 403, 1 492	0.976	0.815
4 wavelengths	1 151, 1 403, 1 492, 1 471	0.980	0.756

Table 1. Result of stepwise regression model

View in the Article

Wavelength	Calibration set		Predicton set
Wavelength	R²	RMSE	R²	RMSE
1 151, 1 403	0.966	1.276	0.960	1.324
1 150, 1 400	0.965	1.301	0.954	1.394

Table 2. Results of modeling and prediction based on feature wavelengths

View in the Article

Bandwidth	Linear regression equation	R²	RMSE
10	y=226.276+292.372x_{1 150}+62.691x_{1 400}	0.961	1.343
25	y=237.591+293.908x_{1 150}+75.411x_{1 400}	0.966	1.253
50	y=275.305+315.264x_{1 150}+85.431x_{1 400}	0.964	1.286
100	y=261.361+269.392x_{1 150}+126.182x_{1 400}	0.961	1.335
200	y=258.446+214.797x_{1 150}+205.532x_{1 400}	0.948	1.548
300	y=204.405-240.874x_{1 150}+762.535x_{1 400}	0.924	1.875

Table 3. Modeling results for different bandwidths

View in the Article

Drift	0	0.01	0.03	0.05	0.07	0.09	0.1	0.2	0.3	0.4	0.5
RMSE	1.253	1.255	1.267	1.292	1.327	1.374	1.400	1.769	2.254	2.79	3.368

Table 4. The effects of central wavelength drift on model prediction errors for 25 nm bandwidth filter centered at 1 150 and 1 400 nm

Download Citation

Save the article for my favorites

Paper Information