In Situ Soft Plant Wearable System for Long-Term and Quantitative Monitoring of Live Plant Stresses

Our work presents a soft plant wearable system (SPWS) that can in-situ, quantitatively, and long-termly monitor various essential abiotic/biotic stresses of live tomato plants, e.g., heat, drought, nutrient deficiencies, and mite attacks. Being intimately attached to the abaxial surface of live tomato leaf, the presented SPWS enables us, for the first time, to simultaneously monitor temperature, humidity, and spectral information of leaf, which can directly indicate the plant stresses. Using a machine learning framework, the SPWS can achieve high prediction accuracy (98.4%) and early-stage diagnostic of stresses (2 days earlier than visual detection). We further applied the SPWS to continuously monitor the plant health fluctuation under stresses, providing a potential solution for noninvasive phenotyping and precision agriculture.<br/>Tomatoes are cultivated worldwide for their naturally healthy and high nutritional value, becoming one of the most produced agricultural products. However, tomato cultivation faces substantial challenges due to various biotic stresses (e.g., diseases and pests) and abiotic stresses (e.g., unsuitable temperature, drought, and nutrient deficiency). These stresses significantly lead to production overcosts, pesticide overuse, and crucial yield losses. Consequently, timely and precise tomato stress monitoring is significant for optimizing pesticide applications and fertilization, improving yield, and protecting the ecosystem. Although some soft sensors can be directly attached to plant surfaces for monitoring their physiological data, a key challenge associated with plant soft sensors is establishing quantitative relationships between sensing data and actual biotic/abiotic stress levels in plants.<br/>In this work, by coupling multi-modal sensing components (e.g., spectral, temperature and humidity (T&H) sensor) with flexible liquid metal circuits, the PSWS can be directly attached to the lower epidermis of tomato leaf for conformal, quantitative, and long-term stresses monitoring. The PSWS introduced a novel in-situ spectral detection method that can collect transmission spectrum of leaves to comprehensively reflect variations in plant pigments. When the plant experiences various stresses, such as essential nutrient deficiency (e.g., nitrogen (N), phosphorus (P), and potassium (K)) and spider mites attackes, leaf pigment composition undergoes significant changes. By analyzing these spectral changes, we can establish a quantitative correlation between specific spectral features and the respective stress factors. Moreover, the PSWS can continuousely detect the leaf surface’s T&H information to indicate transpiration rate and water content of plant for over 10 days. We have also developed a machine learning framework to couple these sensors’ information and classify plant health statuses with high diagnostic accuracy (&gt; 98.4%). The real-time data can be collected and wirelessly transmitted to a mobile user interface, providing instantaneous diagnostic feedback on plant stress conditions..<br/>Finally, two demonstrations of our PSWS have been presented to explore its practical potential for tomato plant cultivation and breeding. 1) Using the PSWS, we continousely monitored health status flactuations of live tomato plants that experienced drought, nutrient deficiency, and timely replenishment for over 31 days. 2) The PSWS can be practically applied in a greenhouse setting, where it was employed to select suitable grafting rootstock by detecting nutritional statuses of various tomato plants.