Raman Spectroscopy and Improved Inception Network for Determination of FHB-Infected Wheat Kernels

Abstract

Detection of infected kernels is important for Fusarium head blight (FHB) prevention and product quality assurance in wheat. In this study, Raman spectroscopy (RS) and deep learning networks were used for the determination of FHB-infected wheat kernels. First, the RS spectra of healthy, mild, and severe infection kernels were measured and spectral changes and band attribution were analyzed. Then, the Inception network was improved by residual and channel attention modules to develop the recognition models of FHB infection. The Inception–attention network produced the best determination with accuracies in training set, validation set, and prediction set of 97.13%, 91.49%, and 93.62%, among all models. The average feature map of the channel clarified the important information in feature extraction, itself required to clarify the decision-making strategy. Overall, RS and the Inception–attention network provide a noninvasive, rapid, and accurate determination of FHB-infected wheat kernels and are expected to be applied to other pathogens or diseases in various crops.

Next‐generation methods for early disease detection in crops

“…132 RS combined with deep learning networks were recently used for the detection of Fusarium head blight (FHB)-infected wheat kernels. 133…”

“…Raman spectroscopy was adopted to detect rot disease of maize caused by Colletotrichum graminicola , 130 rose rosette infection 131 and to investigate wheat and sorghum grains infected with ergot, black tip or mold 132 . RS combined with deep learning networks were recently used for the detection of Fusarium head blight (FHB)‐infected wheat kernels 133 …”

Spectroscopy‐based prediction of 73 wheat quality parameters and insights for practical applications

“…While the NIR spectrometer consistently yielded the best results, the Raman spectrometer was nearly as accurate. However, using a 1064 nm laser instead of a 785 nm laser, as in most Raman and wheat‐related publications (Czaja et al, 2020; Qiu et al, 2022), could reduce the baseline and enhance prediction accuracy. The fluorescence spectrometer used for flour samples was unsuitable for long‐term measurements due to nonlinear changes in the light source spectrum, necessitating its replacement during the study.…”

Diagnosing arsenic-mediated biochemical responses in rice cultivars using Raman spectroscopy

“…This technique is primarily based on the relative concentrations of metabolites present within a plant, namely Raman active molecules phytochemicals such as terpenes, phenolics, and alkaloids. For example, a study in 2022 showed that RS could detect Fusarium head blight in wheat kernels due to the spectral changes in bands associated with lignin, carotenoids, pectin, cellulose, protein, and starch ( Qiu et al., 2022 ). It was also recently demonstrated that RS could be used to diagnose nutritional deficiencies, salinity stress, and aluminum and iron toxicities with high accuracy in rice crops ( Sanchez et al., 2020 ; Higgins et al., 2022a ).…”