AI/ML Driven UAV Based Hyperspectral Imaging for Crop Phenotyping: 

The increasing need for sustainable and climate-resilient agriculture demands rapid, accurate, and scalable crop phenotyping solutions. Our research leverages UAV-based hyperspectral imaging and AI/ML techniques to enable non-invasive monitoring of crop traits, stress responses, and nutrient status at high spatial and spectral resolutions. We develop machine learning/deep learning (ML/DL) algorithms for hyperspectral image enhancement, spectral feature extraction, early water stress detection, and crop nitrogen assessment. The goal is to create efficient end-to-end intelligent systems that support next-generation precision agriculture and crop improvement programs.

Vision-Language Framework for Explainable Hyperspectral Crop Phenotyping: 

This research focuses on developing explainable AI frameworks that combine hyperspectral imaging with vision-language models (VLMs) for agricultural applications. We aim to build domain-aware foundation models that learn rich spectral representations through self-supervised learning and integrate agricultural knowledge with natural language reasoning. Such systems enable agronomists and researchers to interact with hyperspectral data using plain-language queries for tasks such as crop stress detection, growth-stage monitoring, and decision support. The overarching goal is to make hyperspectral AI models more interpretable, trustworthy, and practical for real-world agricultural deployment.