AUTHOR=Tang Zhehan , Jin Yufang , Brown Patrick H. , Park Meerae 

TITLE=Estimation of tomato water status with photochemical reflectance index and machine learning: Assessment from proximal sensors and UAV imagery

JOURNAL=Frontiers in Plant Science

VOLUME=Volume 14 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2023.1057733

DOI=10.3389/fpls.2023.1057733

ISSN=1664-462X

ABSTRACT=Tracking plant water status is a critical step towards the adaptive precision irrigation management of processing tomatoes, one of the most important specialty crops in California. The photochemical reflectance index (PRI) signals from proximal sensors and high-resolution UAV imagery provide an opportunity to monitor crop water status efficiently. Based on data over an experimental tomato field, we developed random forest machine learning regression models to estimate tomato stem water potential (ψ_stem), using observations from proximal sensors and 12-band UAV imagery, respectively, along with weather data. The proximal sensor-based model estimation agreed well with the measured ψ_stem with R2 of 0.74 and MAE of 0.63 Bars, driven by PRI, NDVI, vapor pressure deficit and air temperature. The model tracked well seasonal dynamics of ψ_stem across different plots. A separate model, built with multiple VIs from UAV imagery and weather variables, had an R2 of 0.81 and MAE of 0.67 Bars. The plant-level ψ_stem maps generated from UAV imagery represented well the water status differences over plots under different irrigation treatments, and also tracked well the temporal change among flights. PRI was found as the most important VIs in both proximal sensor- and UAV-based models, providing critical information on tomato plant water status. This study demonstrated that machine learning models can accurately estimate water status levels by integrating PRI, other VIs and weather data, and thus facilitate data-driven irrigation management for processing tomatoes.