Explainable and Lightweight Deep Learning Frameworks for Biotic and Abiotic Stress Assessment in Crops.
Implementing Organization
NERIST, Arunachal Pradesh https://nerist.ac.in/
Principal Investigator
Aswini Kumar Patra
NERIST, Arunachal Pradesh https://nerist.ac.in/
akp@nerist.ac.in
CO-Principal Investigator
Aswini Kumar Patra
NERIST, Arunachal Pradesh https://nerist.ac.in/
Project Overview
This project aims to develop explainable and lightweight deep learning frameworks for detecting biotic (diseases in potato, tomato, cotton, tea) and abiotic stresses (drought and nitrogen deficiency) in crops. We will devise novel deep learning based hybrid models models with explainable AI to interpret stress-related features from plant images. Lightweight architectures and vegetation indices will support efficient field deployment. Few-shot and metric learning will address limited training data, while machine unlearning will enable models to adapt by removing outdated or erroneous data.
Achievements
The project being undertaken by ISRO in collaboration with academia is directly or indirectly beneficial to the society.