An AI & IoT-Enabled Framework for Real-Time Detection of Silent Stress in Crops Using Plant Bio-Signals

Authors

  • K.Thangadurai , S.Ananth , V.Aravindraj , K.Arun Kumar , V.Arunachala Eshwar , R.Logesh , M.Kajendhiran Author

DOI:

https://doi.org/10.7492/4rjppb94

Keywords:

Plant Bio-potentials, Silent Stress Detection, Edge Computing in Agriculture, IoT-Enabled Precision Farming, Bio-signal Classification, Thing Speak Cloud Analytics

Abstract

The agricultural industry across the globe is under great risk due to silent environmental hazards that cannot be detected by human eyes until there is physical
damage. In this regard, the presented research introduces a novel and revolutionary solution for detecting crop silent stress using AI & IoT-Based Crop Silent Stress
Detection System. This technology is based on the plant's internal status and captures bio-potential signals. This is because bio-potential signals occur prior to physical
symptoms. In order to detect changes occurring inside the plant, a highly accurate AD8232 sensor and ESP32 microcontroller were used to collect the bio-potential
signal. The acquired signal will be sent to a cloud-based computing environment. The collected data will be classified into various categories of status including healthy
plants, water stress and nutrient deficiency using Machine Learning algorithm programmed in Python language. Monitoring the plant's status will be done using a
Thing Speak online web platform, which shows the live monitoring results. The experiment showed that bio-signal can detect stress up to 48 hours before any wilt
occurs visually. This work has brought a great combination of traditional plant electrophysiology and the recent IoT design, thereby providing an accurate and
economical way of precision farming, where one can implement preventive measures to improve crop output without any wastage of resources. In addition to data
sensing, the key contribution in this project is the usage of the stacked LSTM network designed particularly for processing the non-linear nature of plant potentials.
The deep learning algorithm provides up to 94.2% accuracy in classifying stress signals while eliminating environmental noise.

Downloads

Published

1990-2026

Issue

Section

Articles

How to Cite

An AI & IoT-Enabled Framework for Real-Time Detection of Silent Stress in Crops Using Plant Bio-Signals. (2026). MSW Management Journal, 36(2), 494-499. https://doi.org/10.7492/4rjppb94