Nondestructive nitrogen content estimation in tomato plant leaves by Vis-NIR hyperspectral imaging and
regression data models

Razieh Pourdarbani; Razieh Pourdarbani; Sajad Sabzi; Mohammad H. Rohban; Ginés García-Mateos; Juan I. Arribas; Juan I. Arribas

doi:10.1364/AO.431886

Applied Optics
Vol. 60,
Issue 30,
pp. 9560-9569
(2021)
•https://doi.org/10.1364/AO.431886

Nondestructive nitrogen content estimation in tomato plant leaves by Vis-NIR hyperspectral imaging and regression data models

Razieh Pourdarbani, Sajad Sabzi, Mohammad H. Rohban, Ginés García-Mateos, and Juan I. Arribas

Not Accessible

Your library or personal account may give you access

Get PDF
Email
Share
Get Citation
Copy Citation Text
Razieh Pourdarbani, Sajad Sabzi, Mohammad H. Rohban, Ginés García-Mateos, and Juan I. Arribas, "Nondestructive nitrogen content estimation in tomato plant leaves by Vis-NIR hyperspectral imaging and regression data models," Appl. Opt. 60, 9560-9569 (2021)

Export Citation
- BibTex
- Endnote (RIS)
- HTML
- Plain Text
Citation alert
Save article

Check for updates

Abstract

The present study aims to estimate nitrogen (N) content in tomato (Solanum lycopersicum L.) plant leaves using optimal hyperspectral imaging data by means of computational intelligence [artificial neural networks and the differential evolution algorithm (ANN-DE), partial least squares regression (PLSR), and convolutional neural network (CNN) regression] to detect potential plant stress to nutrients at early stages. First, pots containing control and treated tomato plants were prepared; three treatments (categories or classes) consisted in the application of an overdose of 30%, 60%, and 90% nitrogen fertilizer, called N-30%, N-60%, N-90%, respectively. Tomato plant leaves were then randomly picked up before and after the application of nitrogen excess and imaged. Leaf images were captured by a hyperspectral camera, and nitrogen content was measured by laboratory ordinary destructive methods. Two approaches were studied: either using all the spectral data in the visible (Vis) and near infrared (NIR) spectral bands, or selecting only the three most effective wavelengths by an optimization algorithm. Regression coefficients (R) were ${0.864}\;{\pm}\;{0.027}$ for ANN-DE, ${0.837}\;{\pm}\;{0.027}$ for PLSR, and ${0.875}\;{\pm}\;{0.026}$ for CNN in the first approach, over the test set. The second approach used different models for each treatment, achieving R values for all the regression methods above 0.96; however, it needs a previous classification stage of the samples in one of the three nitrogen excess classes under consideration.

Full Article | PDF Article

More Like This

Nondestructive detection of anthocyanin content in fresh leaves of purple maize using hyperspectral data

Xin Yang, Shichen Gao, Xiaohe Gu, Chao Zhang, Qian Sun, Zhonghui Wei, Xueqian Hu, and Xuzhou Qu
Appl. Opt. 61(21) 6213-6222 (2022)

Indirect quantitative analysis of soluble solid content in citrus by the leaves using hyperspectral imaging combined with machine learning

Lianbo Guo, Weihua Hunag, Weinan Zheng, Feng Chen, Weiliang Wang, Deng Zhang, Zhenlin Hu, and Yanwu Chu
Appl. Opt. 61(2) 491-497 (2022)

Thermography and machine learning techniques for tomato freshness prediction

Jing Xie, Sheng-Jen Hsieh, Hong-Jin Wang, and Zuojun Tan
Appl. Opt. 55(34) D131-D139 (2016)

Previous Article Next Article

Data Availability

Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.

Cited By

You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Figures (6)

You do not have subscription access to this journal. Figure files are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Tables (4)

You do not have subscription access to this journal. Article tables are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Equations (2)

You do not have subscription access to this journal. Equations are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Layer Type	Convolution Size	Output Shape ( $Width \times Features$ )	Number of Params.
$C o n v o l u t i o n + R e L u$	$7 \times 1$	$321 \times 64$	512
Max pooling		$160 \times 64$	–
$C o n v o l u t i o n + R e L u$	$7 \times 64$	$154 \times 64$	28,736
Max pooling		$77 \times 64$	–
$C o n v o l u t i o n + R e L u$	$5 \times 64$	$73 \times 128$	41,088
Max pooling		$36 \times 128$	–
$C o n v o l u t i o n + R e L u$	$5 \times 128$	$32 \times 128$	82,048
Max pooling		$16 \times 128$	–
$C o n v o l u t i o n + R e L u$	$5 \times 128$	$12 \times 256$	164,096
Max pooling		$6 \times 256$	–
$C o n v o l u t i o n + R e L u$	$3 \times 256$	$5 \times 512$	393,728
Flatten		2560	–
Dense		1	2,561

Method	MSE	RMSE	MAE	R	$R^{2}$
ANN-DE	$0.2413 \pm 0.0404$	$0.4896 \pm 0.0405$	$0.3784 \pm 0.0277$	$0.8638 \pm 0.0267$	$0.7469 \pm 0.0458$
PLSR	$0.2919 \pm 0.0519$	$0.5382 \pm 0.0474$	$0.4149 \pm 0.0289$	$0.8371 \pm 0.0271$	$0.7015 \pm 0.0454$
CNN	$0.2216 \pm 0.0415$	$0.4687 \pm 0.0438$	$0.3516 \pm 0.0305$	$0.8749 \pm 0.0257$	$0.7660 \pm 0.0447$

Method	N Excess (%)	MSE	RMSE	MAE	R	$R^{2}$
ANN-DE	30	$0.0049 \pm 0.0107$	$0.0577 \pm 0.0397$	$0.0415 \pm 0.0228$	$0.9662 \pm 0.0541$	$0.9365 \pm 0.0865$
	60	$0.0064 \pm 0.0058$	$0.0753 \pm 0.0274$	$0.0522 \pm 0.0144$	$0.9636 \pm 0.0303$	$0.9293 \pm 0.0562$
	90	$0.0002 \pm 0.0002$	$0.0146 \pm 0.0051$	$0.0112 \pm 0.0036$	$0.9744 \pm 0.0248$	$0.9502 \pm 0.0457$
PLSR	30	$0.0013 \pm 0.0020$	$0.0312 \pm 0.0171$	$0.0223 \pm 0.0100$	$0.9602 \pm 0.0554$	$0.9251 \pm 0.0936$
	60	$0.0037 \pm 0.0016$	$0.0602 \pm 0.0118$	$0.0437 \pm 0.0069$	$0.9806 \pm 0.0077$	$0.9616 \pm 0.0151$
	90	$0.0002 \pm 0.0000$	$0.0126 \pm 0.0017$	$0.0098 \pm 0.0013$	$0.9837 \pm 0.0039$	$0.9677 \pm 0.0076$

Layer Type	Convolution Size	Output Shape ( $Width \times Features$ )	Number of Params.
$C o n v o l u t i o n + R e L u$	$7 \times 1$	$321 \times 64$	512
Max pooling		$160 \times 64$	–
$C o n v o l u t i o n + R e L u$	$7 \times 64$	$154 \times 64$	28,736
Max pooling		$77 \times 64$	–
$C o n v o l u t i o n + R e L u$	$5 \times 64$	$73 \times 128$	41,088
Max pooling		$36 \times 128$	–
$C o n v o l u t i o n + R e L u$	$5 \times 128$	$32 \times 128$	82,048
Max pooling		$16 \times 128$	–
$C o n v o l u t i o n + R e L u$	$5 \times 128$	$12 \times 256$	164,096
Max pooling		$6 \times 256$	–
$C o n v o l u t i o n + R e L u$	$3 \times 256$	$5 \times 512$	393,728
Flatten		2560	–
Dense		1	2,561

Method	MSE	RMSE	MAE	R	$R^{2}$
ANN-DE	$0.2413 \pm 0.0404$	$0.4896 \pm 0.0405$	$0.3784 \pm 0.0277$	$0.8638 \pm 0.0267$	$0.7469 \pm 0.0458$
PLSR	$0.2919 \pm 0.0519$	$0.5382 \pm 0.0474$	$0.4149 \pm 0.0289$	$0.8371 \pm 0.0271$	$0.7015 \pm 0.0454$
CNN	$0.2216 \pm 0.0415$	$0.4687 \pm 0.0438$	$0.3516 \pm 0.0305$	$0.8749 \pm 0.0257$	$0.7660 \pm 0.0447$

Abstract

Data Availability

Cited By

Figures (6)

Tables (4)

Equations (2)

Applied Optics