Monitoring Nitrogen Nutrition and Grain Protein Content of Rice Based on Ensemble Learning

Jie ZHANG; Bo XU; Hai-kuan FENG; Xia JING; Jiao-jiao WANG; Shi-kang MING; You-qiang FU; Xiao-yu SONG

doi:10.3964/j.issn.1000-0593(2022)06-1956-09

Spectroscopy and Spectral Analysis, Volume. 42, Issue 6, 1956(2022)

Monitoring Nitrogen Nutrition and Grain Protein Content of Rice Based on Ensemble Learning

Jie ZHANG1,1; 2;... Bo XU1,1;, Hai-kuan FENG1,1;, Xia JING2,2;, Jiao-jiao WANG1,1;, Shi-kang MING1,1;, You-qiang FU3,3; and Xiao-yu SONG1,1; *; |Show fewer author(s)

Author Affiliations

¹1. Information Technology Research Center, Beijing Academy of Agriculture and Forestry Sciences, Beijing 100094, China

²2. School of Surveying and Mapping Science and Technology, Xi'an University of Science and Technology, Xi'an 710054, China

³3. Rice Research Institute, Guangdong Academy of Agricultural Sciences, Guangzhou 510640, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(9)

Fig. 1. Correlation between canopy spectrum and nitrogen parameters at different growth stages of rice in 2019 and 2020

Download full size

View in Article

Fig. 2. R², RMSE and MAE based on the canopy spectral data seed protein content model of rice at different fertility stages

Download full size

View in Article

Fig. 3. R², RMSE and MAE of seven algorithms under different parameter combinations at Panicle Initiation and Heading stage

Download full size

View in Article

Fig. 4. R², RMSE and MAE of seven algorithms with full band spectral information and PNA as input in Panicle Initiation stage

Download full size

View in Article

Fig. 5. R², RMSE and MAE of seven algorithms with full band spectral information and PNC as input at Heading stage

Download full size

View in Article

Table 1. Test data acquisition
View table
View in Article
Table 1. Test data acquisition
参数 2019年 2020年
分化期抽穗期分化期抽穗期
冠层光谱 √ √ √ √
长势参数(LNC, LNA,
PNC, PNA) √ √ √ √
水稻品质数据 √ √ √ √

Table 2. Model accuracy of nitrogen parameters based on canopy spectral data of rice at Panicle Initiation stage

View table

View in Article

Table 2. Model accuracy of nitrogen parameters based on canopy spectral data of rice at Panicle Initiation stage

方法	LNC/%			LNA/(g·m^-2)			PNC/%			PNA/(g·m^-2)
方法	R²	RMSE	MAE	R²	RMSE	MAE	R²	RMSE	MAE	R²	RMSE	MAE
PLSR	0.283	0.284	0.198	0.581	0.648	0.477	0.319	0.197	0.131	0.608	0.913	0.595
KNN	0.527	0.232	0.155	0.609	0.627	0.411	0.490	0.174	0.097	0.613	0.911	0.562
BRR	0.418	0.258	0.097	0.678	0.569	0.217	0.571	0.158	0.071	0.767	0.706	0.323
SVR	0.606	0.213	0.096	0.697	0.556	0.259	0.712	0.130	0.070	0.772	0.708	0.352
RF	0.912	0.120	0.110	0.944	0.272	0.230	0.930	0.084	0.079	0.938	0.399	0.353
Adaboost	0.927	0.110	0.187	0.932	0.304	0.504	0.920	0.082	0.144	0.935	0.402	0.734
Bagging	0.807	0.154	0.219	0.915	0.311	0.532	0.861	0.103	0.153	0.914	0.465	0.723

Table 3. Model accuracy of nitrogen parameters based on canopy spectral data of rice heading stage

View table

View in Article

Table 3. Model accuracy of nitrogen parameters based on canopy spectral data of rice heading stage

方法	LNC/%			LNA/(g·m^-2)			PNC/%			PNA/(g·m^-2)
方法	R²	RMSE	MAE	R²	RMSE	MAE	R²	RMSE	MAE	R²	RMSE	MAE
PLSR	0.699	0.202	0.156	0.645	1.081	0.834	0.705	0.110	0.080	0.619	2.256	1.727
KNN	0.630	0.199	0.141	0.677	1.084	0.740	0.679	0.098	0.083	0.612	2.354	1.696
BRR	0.738	0.112	0.084	0.658	0.504	0.385	0.779	0.061	0.049	0.652	1.221	0.943
SVR	0.749	0.104	0.089	0.682	0.506	0.410	0.827	0.063	0.055	0.626	1.216	1.071
RF	0.944	0.137	0.103	0.948	0.591	0.436	0.951	0.073	0.057	0.920	1.346	1.042
Adaboost	0.954	0.241	0.185	0.937	1.057	0.826	0.938	0.133	0.109	0.909	2.387	1.918
Bagging	0.894	0.216	0.159	0.905	1.100	0.833	0.911	0.127	0.093	0.888	2.358	1.792

Table 4. Model accuracy of nitrogen parameters based on canopy spectral data of rice whole growth period

View table

View in Article

Table 4. Model accuracy of nitrogen parameters based on canopy spectral data of rice whole growth period

方法	LNC/%			LNA/(g·m^-2)			PNC/%			PNA/(g·m^-2)
方法	R²	RMSE	MAE	R²	RMSE	MAE	R²	RMSE	MAE	R²	RMSE	MAE
PLSR	0.379	0.309	0.236	0.580	1.028	0.760	0.438	0.267	0.212	0.592	2.553	1.953
KNN	0.375	0.313	0.245	0.639	0.962	0.704	0.332	0.293	0.231	0.584	2.585	1.921
BRR	0.596	0.250	0.198	0.731	0.824	0.622	0.833	0.146	0.116	0.837	1.616	1.202
SVR	0.601	0.250	0.190	0.711	0.868	0.596	0.736	0.186	0.149	0.781	1.900	1.304
RF	0.922	0.135	0.107	0.943	0.440	0.325	0.925	0.115	0.091	0.952	0.978	0.716
Adaboost	0.778	0.199	0.181	0.876	0.596	0.509	0.848	0.154	0.133	0.919	1.426	1.213
Bagging	0.896	0.139	0.105	0.905	0.531	0.376	0.896	0.123	0.094	0.935	1.073	0.789

Tools

Get Citation

Copy Citation Text

Jie ZHANG, Bo XU, Hai-kuan FENG, Xia JING, Jiao-jiao WANG, Shi-kang MING, You-qiang FU, Xiao-yu SONG. Monitoring Nitrogen Nutrition and Grain Protein Content of Rice Based on Ensemble Learning[J]. Spectroscopy and Spectral Analysis, 2022, 42(6): 1956

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Research Articles

Received: Oct. 27, 2021

Accepted: --

Published Online: Nov. 14, 2022

The Author Email:

DOI:10.3964/j.issn.1000-0593(2022)06-1956-09

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Test data acquisition

Table 1. Test data acquisition

Table 2. Model accuracy of nitrogen parameters based on canopy spectral data of rice at Panicle Initiation stage

Table 2. Model accuracy of nitrogen parameters based on canopy spectral data of rice at Panicle Initiation stage

Table 3. Model accuracy of nitrogen parameters based on canopy spectral data of rice heading stage

Table 3. Model accuracy of nitrogen parameters based on canopy spectral data of rice heading stage

Table 4. Model accuracy of nitrogen parameters based on canopy spectral data of rice whole growth period

Table 4. Model accuracy of nitrogen parameters based on canopy spectral data of rice whole growth period

微信扫一扫：分享