4.7 Article

Spectral monitoring of salinity stress in tomato plants

期刊

BIOSYSTEMS ENGINEERING
卷 217, 期 -, 页码 26-40

出版社

ACADEMIC PRESS INC ELSEVIER SCIENCE
DOI: 10.1016/j.biosystemseng.2022.02.018

关键词

non-invasive; machine learning; spectroscopy; Solanum lycopersicum L.; vegetation stress

资金

  1. Ministry of Agriculture and Rural Development, Israel (Eugene Kandel Knowledge Centres) as part of the Root of the Matter project: The Root Zone Knowledge Centre for Leveraging Modern Agriculture [391/15/16-34-0005]

向作者/读者索取更多资源

Water salinity is a common agricultural hazard that affects a significant amount of irrigated land, resulting in reduced crop yield. Although stress coping mechanisms in plants have been studied, there is still a lack of understanding of plant adaptation and acclimation, which often varies among species and varieties. Current methods for assessing plant stress are expensive, destructive, and time-consuming. Spectroscopy, on the other hand, has the potential to monitor biochemical components and physiological states of plants in a non-destructive manner. The goal of this study was to develop a spectral-based model for detecting plants under salt stress, aiming to optimize plant-status monitoring without damaging the plants.
Water salinity is a widespread agricultural hazard that affects approximately 20% of irri-gated land, causing a significant yield reduction in crops. Stress coping mechanisms by plants were thoroughly examined but understanding of plant adaptation and acclimation is still lacking and is often species-and variety-specific. Presently, the biochemical and physiological methods that are used to assess plant stress are costly, destructive, and time-consuming. Alternatively, spectroscopy is a potential method to monitor biochemical components and physiological states of plants. The objective of the current work was to build a spectral-based model for detecting plants under salt stress, in order to optimise plant-status monitoring in a non-destructive manner. In this study, five different tomato graft combinations were examined under four different salinity treatments in a green-house. Hyperspectral measurements were conducted in the range of 400-2500 nm, and chemometrics was used for data analysis and modelling. Salt treatments were found to affect the physiological performance of plants, although environmental conditions had a greater influence on plant temporal physiological trends. Spectral data acquisition with chemometrics showed high ability to predict salt accumulation in plants (root mean square error of prediction (RMSEP) of 0.47 mg g(-1) and 2.8 mg g(-1) for Na+ and Cl-, respectively). Moreover, a hyperspectral, robust decision-supporting classification model was established for detecting plants under salt stress (prediction specificity: 0.94). The presented capabilities of predicting Cl-, Na+, and the K:Na ratio in a non-destructive manner, by utilising spectroscopy, could serve as the basis for developing a low-cost, fast, and efficient stress detection method, independent of environmental conditions. (C) 2022 IAgrE. Published by Elsevier Ltd. All rights reserved.

作者

我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。

评论

主要评分

4.7
评分不足

次要评分

新颖性
-
重要性
-
科学严谨性
-
评价这篇论文

推荐

暂无数据
暂无数据