Internet of Plants: Machine Learning System for Bioimpedance-Based Plant Monitoring

Abstract

Sensors in plant and crop monitoring play a key role in improving agricultural efficiency by enabling the collection of data on environmental conditions, soil moisture, temperature, sunlight, and nutrient levels. Traditionally, wide-scale wireless sensor networks (WSNs) gather this information in real-time, supporting the optimization of cultivation processes and plant management. Our paper proposes a novel “plant-to-machine” interface, which uses a plant-based biosensor as a primary data source. This model allows for direct monitoring of the plant’s physiological parameters and environmental interactions via Electrical Impedance Spectroscopy (EIS), aiming to reduce the reliance on extensive sensor networks. We present simple data-gathering hardware, a non-invasive single-wire connection, and a machine learning-based framework that supports the automatic analysis and interpretation of collected data. This approach seeks to simplify monitoring infrastructure and decrease the cost of digitizing crop monitoring. Preliminary results demonstrate the feasibility of the proposed model in monitoring plant responses to sunlight exposure.

Code and data availability

The supplied article blocks describe a bioimpedance-based plant monitoring system with LSTM classification, but contain no data availability statement, code repository, public dataset link, or trained model release. No paper-specific public asset is identified.

No evidence-backed public reproduction asset is currently recorded.