Luca Brillante,* Eve Laroche-Pinel, Brent Sams, Benjamin Corrales, Kaylah Vasquez, and Vincenzo Cianciola
*Department of Viticulture and Enology, California State University Fresno, 2360 E Barstow Ave, Fresno, CA, 93740 (lucabrillante@csufresno.edu)

Assessing grape composition is essential in vineyard management to decide harvest date and to optimize cultural practices toward achieving production goals. Grape composition is variable in time and space, as it is affected by the ripening process and depends on soil and climate conditions. Our work focused on developing a system to assess and map grape composition directly in the field. For this study, a UTV was specially adapted to lift the canopy and expose the fruits, and two hyperspectral cameras covering the 500 to 1700 nm range were mounted with GPS systems and halogen lights for night imaging. We imaged a Merlot vineyard located in Madera, California during the 2022 growing season. At the same time, we sampled grapes from 160 vine locations that were analyzed in the laboratory for total soluble solids, titratable acidity, pH, and anthocyanin profile. About 1000 samples were collected. For the analysis, the images were segmented to extract the grape’s signal from sampled vines. Then, the reflectance of the grapes was used to look for correlations with grape composition using machine learning models. Evaluation of models was performed using RMSE, and R2 in k-fold cross-validation and through the use of hold-out test sets. The interpretation of the model was conducted through feature importance and partial dependence plots to understand the relationship between wavelength predictors and outcome. This project is the first to use a SWIR camera mounted on a UTV to assess grape composition. Our results demonstrate that SWIR images can be used to perform a classification to extract the grape signal with a mean error of 2.2% using the spectral signature of each class represented in the image (grape, leaves, and background). The prediction of grape compounds from the refined spectral signal shows promising results.

Funding Support: American Vineyard Foundation; California State University – Agricultural Research Institute