Exploring the potential of hyperspectral imaging to detect Esca disease complex in asymptomatic grapevine leaves

• A pixel-based classification of esca-affected grapevine leaves was achieved by HSI. Multiple pre-processing and variable selection methods were tested using PLS-DA. Feature wavelengths identified automatic interval PLS methods. Asymptomatic belonging to vines successfully identified. NIR-HSI could detect foliar symptoms Esca disease complex. Precise reliable identification specific plant diseases is a challenge within precision agriculture nowadays. This the case esca, complex trunk disease, that represents major threat modern viticulture as it responsible for large economic losses annually. The lack effective control strategies complexity esca expression make essential affected plants, before become evident, better management vineyard. study evaluated suitability near-infrared hyperspectral imaging (HSI) system in asymptomatic Tempranillo red-berried cultivar. For this, 72 from an experimental vineyard, naturally infected with collected scanned lab-scale HSI 900-1700 nm spectral range. Then, image processing multivariate analysis techniques merged develop models distinction healthy, symptomatic leaves. Automatic partial least squares identify most relevant detection discriminant different techniques. Three-class two-class classifiers carried out differentiate leaf pixels, healthy respectively. Both performed similarly, achieving good rates range 82.77-97.17% validation datasets either three-class or classifiers. latter results demonstrated capability distinguish two groups seemingly identical (healthy asymptomatic). These findings would ease annual monitoring incidence vineyard and, therefore, crop decision making.