Smart irrigation for management of processing tomato: a machine learning approach

Abstract

Implementation of the Internet of Things (IoT) is revolutionizing agricultural techniques in modern precision irrigation systems. This research was conducted in an open field trial on processing tomato to evaluate the application of machine learning (ML)-based decision support systems (DSS) for predictive optimization of deficit irrigation strategies. A smart irrigation system (SIS) was adopted to investigate the effect of different management strategies under water deficit conditions. Four irrigation treatments using drip irrigation were tested: Full irrigation (FI) - (100% of ETc), Deficit irrigation (DI 1 ) - (80% of ETc), Regulated Deficit irrigation (RDI) - (60-80-60% of ETc) and Deficit irrigation (DI 2 ) - (60% of ETc). The trial was monitored via an ML-based SIS connected to a network of soil moisture and temperature sensors. This study aims to: (i) assess how processing tomato responds to controlled deficit irrigation using an ML-based DSS; (ii) examine the correlation between various water productivity indices (WP, TYWP, MYWP, and IWP) and tomato yields and quality; (iii) investigate the impact of temperature and thermal stress on tomato performances under different irrigation regimes. Results show significant effects of deficit irrigation on growth and yield parameters ( p ≤ 0.001). Both RDI and DI 1 achieved high water productivity, with no significant differences compared to FI. In the 2022 growing season, deficit irrigation, particularly DI 1 , resulted in higher WP values, saving 32.66% of water and improving fruit quality. The study emphasizes the potential of deficit irrigation for achieving high yields and quality in processing tomatoes. However, the ML-based DSS, while effective in water management, requires enhanced sensitivity to crop susceptibility to heat and water stresses.

Topics & Concepts