Developing Fluorescence Hyperspectral Imaging Models for Non-Invasive Detection of Herbicide Safeners Action Mechanism in Wheat Crop

Abstract

Herbicide safeners are considered key agents for plant protection that reduce the harmful impacts of herbicides on crops and the environment in general. Where traditional evaluation methods for their effectiveness are time-consuming. In this study, a rapid and non-destructive method was proposed using chlorophyll fluorescence spectrometry that combined with machine learning models. Besides, chemometric analysis was utilized to reveal the action mechanism between the wheat crop (Triticum aestivum L.) understudy and the herbicide isoproturon (ISO) and safener gibberellin acid (GA3). The results showed that ISO caused oxidative stress and disrupted the photosynthesis mechanism in wheat by hindering the electron transport pathway from primary acceptor quinone (QA) to secondary acceptor (QB). Meanwhile, GA3 stimulated wheat to synthesize more glutathione (GSH) that accelerated the herbicide action metabolism. It’s worth noting that excessive GA3 has decreased significantly the GSH and photosynthetic pigment concentrations, while the malondialdehyde (MDA) concentration was significantly (p < 0.05) increased. Additionally, competitive adaptive reweighted sampling (CARS) proved the best performance when combined with partial least square regression (PLSR) for predicting the phytochemical concentrations that characterized the effectiveness of GA3. In conclusion, the novelty of the current study came from the accurate real-time tracking method for GA3 action mechanism and its effectiveness on ISO toxicity. Where, that model holds great value for reducing the traditional methods’ limitations in safener developments.

Keywords: Visible/near-infrared hyperspectral imaging, chlorophyll a fluorescence, chemometric analysis, machine learning models