Smart Agricultural Technology (Aug 2023)
PLANTSENS: A rail-based multi-sensor imaging system for redundant water stress detection in greenhouses
Abstract
The production of fruits and vegetables in a protected environment (e.g., greenhouse) and open field mainly depends on a substantial water supply. Water stress is the most common plant stress, significantly affecting the quality and yield of crops, especially in organic plant cultivation and arid regions. An automated and well-scheduled irrigation system that corresponds to the actual water demand helps reduce the consumption of this limited resource without negatively affecting crop productivity. This study outlines the design, construction, and functionality of an autonomous multi-sensor monitoring system for real-time water stress detection and precisely targeted irrigation called PLANTSENS. The developed prototype combines the absorption-based and the temperature-based measuring method by using three optical sensors operating in the near-infrared, shortwave, and longwave (thermal) infrared spectrum. The camera modules are mounted on a moving rail-based platform with an integrated positioning system. Appropriate networking and cloud environment with a central server for storing, analyzing, and visualizing the acquired data was created. PLANTSENS monitoring system has been successfully implemented and trained on tomato plants in a greenhouse. The algorithm for scheduling irrigation time has been developed by merging two water stress detection algorithms, namely the Water Index (WI) and the Crop Water Stress Index (CWSI), thereby improving the system's capability and accuracy. WI determines the plant stress from the reflection differences at 1450 and 1300 nm, where water absorbs strongly and weakly, respectively. CWSI, on the other hand, uses the leaf temperature, which tends to increase during water stress due to the lower transpiration rate caused by closed stomata. This work presents the performance of investigated measuring methods and developed algorithms based on the system approach from July 2020 through August 2020. The obtained irrigation time has been verified in accordance with observed changes in the plant's physiological response, such as vapor conductivity, validating this approach as promising and confirming water stress as a significant indicator in irrigation scheduling.
