Remote Monitoring of Plant Water Stress with RGB Imaging

Precision irrigation in greenhouses necessitates remote monitoring of soil moisture. Traditional methods often rely on point measurements, making comprehensive water stress assessment across all crop plants impractical. As an alternative, machine vision has emerged as a promising solution. This study presents a novel approach to soil moisture monitoring using plant images, implementable with low-cost devices and minimal computational resources. The method is based on the hypothesis that leaf discoloration serves as an early indicator of water stress, detectable through RGB imaging. We detail the development and installation of a monitoring system within a grow tent, designed to test irrigation automation based on leaf color across various crops in a controlled environment.