Skip to content

Early Detection of Water Stress by Plant Electrophysiology: Machine Learning for Irrigation Management

Home / Papers / Early Detection of Water Stress by Plant Electrophysiology: Machine Learning for Irrigation Management

Early Detection of Water Stress by Plant Electrophysiology: Machine Learning for Irrigation Management

Authors: Eduard Buss, Till Aust, Heiko Hamann Date: 2026-04-30 Paper ID: openalex:2604.28038

Summary

This paper presents a machine learning framework for the early detection of water stress in greenhouse-grown tomato plants using electrophysiological time-series data. The authors establish an automated processing pipeline that incorporates statistical feature selection and probability calibration, identifying an optimal 30-minute observation window for stress classification. The proposed approach achieves up to 92% accuracy, outperforming deep learning alternatives and enabling effective state-transition detection for autonomous irrigation management.

Key Contributions

  • Developed an electrophysiological signal processing framework for early detection of plant water stress using automated machine learning.
  • Identified a 30-minute look-back window as the optimal trade-off between decision latency and classification accuracy (up to 92%).
  • Demonstrated superior performance of automated machine learning models over deep learning architectures in the context of plant electrophysiology data.

Metadata & Links

url
https://arxiv.org/abs/2604.28038
paper_id
2604.28038
paper_source
openalex
domain
time-series
tags
architectures
datasets
concept_slugs
dataset_slugs
skill
TimeSeriesSkill
created_at
2026-05-03T06:03:41Z