LSTM Neural Networks Anomaly Detection for Biotic and Abiotic Early Stress Detection on Tomato Plants

This study investigates time-series anomaly detection using a Long Short-Term Memory (LSTM) neural network to identify early biotic and abiotic stress conditions in tomato plants. The approach analyzes a stem frequency parameter related to stem conductivity, which reflects plant hydration and physiological status. An LSTM model trained on healthy plant data was used to predict future stem frequency values, and anomalies were detected when predictions significantly deviated from measured data. Experimental results show that the proposed method can detect water stress several days before visible symptoms appear and, in most cases, provides early warnings of biotic stress. The study highlights the potential of LSTM-based anomaly detection for early plant stress monitoring in precision agriculture.