S06 - Session P9 - Early crop stress detection using soft-sensors

Authors: Laura Cammarisano, Mehdi Bisbis *, Leo Marcelis, Ep Heuvelink, Oliver Körner

In horticultural production, environmental stresses such as high/low temperature, salinity, drought, high/low nutrient supply or excessive radiation are among the main obstacles in reaching the whole yield potential. A crop monitoring system for early stress detection, i.e. a warning tool at a stage of reversibility of stress, would among others enable dynamic climate control and increase resource use efficiency in climate controlled cultivation systems as greenhouses and indoor farms. We reviewed the state of the art in soft-sensor research and application and present an innovative concept of a crop monitoring system for the detection of early stress signals in tomato. We propose a bouquet of smart combinations of physical sensors with mathematical models as a basis for real-time monitoring and decision support system (DSS). In our proposed DSS-design, leaf net-photosynthesis (P n ), stomatal conductance (g s ), leaf temperature (T l ) as well as the quantum yield of photosystem II (ϕ PSII ) are used as signal variables. The difference of model-calculated and measured time courses is used as an indicator of a plant's stress level. Models serve as baseline providers in a dynamic environment; i.e., the model calculates the optimum range of the mentioned physiological variables under the prevailing dynamically changing environmental conditions. Sensor readings outside that range indicate the probability of a stressed plant or crop. Furthermore, we provide a simulation study with various scenarios to show that with our approach dynamic greenhouse set-points can be applied without the risk for stress related crop damages, yielding high energy savings.