Low-Cost Handheld Spectrometry for Detecting Flavescence Dorée in Vineyards

This study was conducted to evaluate the potential of low-cost hyperspectral sensors for early detection Flavescence dorée (FD) from asymptomatic samples prior symptom development. In total, 180 leaf spectra 60 randomly selected plants (three leaves per plant) were collected by using two portable mini-spectrometers (Hamamatsu: 340–850 nm and NIRScan: 900–1700 nm) at five vegetative growth stages in a vineyard with grape variety Garganega. High differences Hamamatsu groups found VIS-NIR (visible–near infrared) spectral region while very small observed NIRScan spectra. We analyzed data all bands, features reduced an ensemble method, genetic algorithms (GA) discriminate healthy (FD negative) diseased positive) different classifiers. Overall, high classification accuracies case sensor compared sensor. The feature selection techniques performed better highest accuracy 96% achieved when GA used logistic regression (LR) classifier on test samples. A slightly low 85% (selected method) support vector machine (SVM) leave-one-out (LOO) cross-validation whole dataset. Results demonstrated that employing technique can provide valid tool determining optimal bands be identify FD disease vineyard. However, further validation studies are required, as this dataset single grapevine variety.