AI-Enabled Agricultural Evaluation of Seaweed Biostimulants for Tomato Growth, Yield, and Quality

Abstract

Seaweed-derived biostimulants are increasingly investigated as sustainable crop inputs capable of improving plant development, productivity, fruit quality, and tolerance to environmental stress.
However, conventional evaluations generally depend on periodic destructive measurements and end-of-season observations, providing limited information on the temporal and nonlinear responses of tomato plants to different biostimulant concentrations. This study proposes an AI-enabled agricultural evaluation framework that combines longitudinal RGB–multispectral plant imaging, physiological measurements, environmental observations, and conventional agronomic traits for evaluating seaweed biostimulant responses in tomato (Solanum lycopersicum L.). The central research gap concerns the absence of integrated predictive systems capable of linking early phenotypic responses to subsequent yield and fruit-quality outcomes while simultaneously identifying an agronomically optimal biostimulant dose. A multimodal learning strategy is conceptualized to extract canopy-growth signatures, vegetation characteristics, flowering dynamics, fruit-development traits, and environmental interactions across untreated and seaweed-treated plants.