Highly Stretchable and Reliable Graphene-Based Strain Sensor for Plant Health Monitoring and Deep Learning-Assisted Crop Recognition

Abstract

Stretchable sensors hold great potential for monitoring plant physiological parameters and enabling crop identification in smart agriculture. However, achieving long-term, stable, reliable monitoring of plants in dynamic environments, as well as improving crop identification accuracy, remains a substantial challenge, primarily due to the limited biocompatibility of conventional stretchable sensors. Here, we present a highly stretchable and reliable strain sensor based on a graphene/Ecoflex composite. This sensor features a mesh structure that combines graphene's high electrical conductivity and strain sensitivity with Ecoflex's excellent stretchability, biocompatibility, and resistance to environmental degradation. By structural optimization, the sensor achieves high sensitivity (gauge factor = 138), a low detection limit (0.1% strain), and high reliability (over 1,500 cycles), along with waterproofing and resistance to both acidic and alkaline conditions. Furthermore, the sensor conforms tightly to various plant leaves and stems without hindering growth, enabling real-time monitoring of plant growth patterns and in situ detection of mechanical damage to predict plant stress. Moreover, assisted by deep learning, it precisely classifies 8 crop types with an accuracy of 95.2%. These demonstrate that stretchable sensors based on mesh graphene/Ecoflex can operate reliably in outdoor agricultural environments even in the face of variable climatic and chemical conditions, providing a practical platform for advancing plant phenomics and smart agricultural robotics.