IoT and Deep Learning Detect Termite Infestations in Tea Plantations

Summary

Researchers developed an IoT-enabled acoustic monitoring framework that uses a convolutional neural network to detect Upcountry Live Wood Termite infestations in tea plantations. A Raspberry Pi and high-sensitivity microphone captured non-invasive trunk sounds, while coordinates enabled spatial tracking. The CNN trained on spectrograms reached 81.5% accuracy and 0.819 ROC-AUC on a held-out test set. A weighted model combined infestation probability, acoustic amplitude, and nearby affected plants to estimate severity and identify high-risk areas for targeted inspections and control.